(Posted to Bob and to edstat)

On Thu, 19 Feb 2004, Robert Frick wrote in part:

> ... I was thinking ... that total was the easy concept and average the
> difficult concept. But there are too many natural examples where
> people are talking about average, not total. When we say someone is
> usually happy, we are not making a claim about how many times we have
> seem them happy, but instead a claim about what percentage of the time
> they are happy.

Well, I'm not so sure.  It strikes me that "a claim about what
percentage of the time they are happy" is the way a person with some
arithmetical skills might choose to characterize (or operationalize) the
idea of "usually happy", but I do not see "percentage" (or any other
representation of a fraction) as inherent, nor even necessarily correct.
The perception of "usually happy" might well arise from behavior that is
vivid, vigorous, and expressive when the person in question is evidently
happy, coupled with behavior that is subdued, quiet, perhaps even absent
from view when the person is unhappy;  so that the "averaging" resulting
in the judgement "usually happy" may not be the proportion of time the
person IS happy, but more like the proportion of time that I _see_ him
when he's happy, and possibly even that proportion weighted by an
intensity factor (actually, he's happy only about 1/3 of the time I see
him, but he's so vividly happy that I'm unaware that he's unhappy twice
as frequently).  And "percentage"  probably should be interpreted rather
broadly:  I'd not be surprised if the precision of one's perception was
a good deal less precise than one part in a hundred.
 -- Don.
 ------------------------------------------------------------
 Donald F. Burrill                              [EMAIL PROTECTED]
 56 Sebbins Pond Drive, Bedford, NH 03110      (603) 626-0816
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to