Flickering bulb #2

["John Constantine" <thesfg1@cox.net>]

Picking up where I left off, in regard to considerations enumerative vs.
analytical - problems arise from a number of factors and sources which may
lead an administrator of some persuasion (public, private, operational,
educational) to assume that person A, who went to college A1 will be a
better employee that person B who went to college B2. An employer might take
a point of view, unfortunately, that person E, having obtained a grade of
3.9 will make a better employee than will person K, who obtained a grade of
1.9.

Why are these problematic? Mixing of Apples and kiwifruit? Right idea, wrong
approach? Bad data? All of the above, and more?

As Deming pointed out in TNE, just because you know (or think you know)
something about an item from your enumerative study of a sample does not
mean to imply that you could project forward in time (predict) what that
item, or member of that sample, will do in the next minute much less into
the future months and years hence. Yet how many times have we seen this
done, via reports and surveys which seek to do just that. While you could
say a great deal about a sample, even to the point of 99% confidence, having
done so your job is, so to speak, over. Thus spake Zarathustra.

If, by some chance the administrator pointed to a chart and voiced an
opinion and the actual result were "close enough for use", fine. But, if the
opinion or prediction was not in the ball park, the control chart would
surely stand out in bold print as an indicator of the value of the
analytical, inferential approach to prediction. Yet, how many know of it
and/or utilize it?

Lessons learned? Perhaps one might be to best avoid sampling, ranking, and
rating if the intent is to look for better items or better candidates among
the sample. Far, far better to use the real data, historical data, and
empirical data to infer what might be happening and the variation within the
process(es).

Do others see merit in the distinction between enumerative and analytical
studies?

Regards,

________________
John Constantine
thesfg1@cox.net
Phoenix, AZ