At around 6/8/06 11:07 am, Leigh Meyers wrote:
>> Or a salesman tells you that the median commission on the nine sales
>> he made that week is $80 and suggests that he therefore made $720 on
>> these sales. Maybe, but he could have made millions of dollars on
>> these sales if one of the sales was enormous, or he might have made
>> little more than $400 if four of the sales were near $0 and the other
>> five were $80 or slightly more.
>
> Like I said, the median means squat,. or, as raghu suggested, it coud be
> "certainly greater than "squat".", or anywhere in between squat and not...
>
> Weasel words and Nothing numbers. It's what makes America Illusorily great.
>
Leigh,
I am afraid I still do not get your criticism. It seems that you are
questioning the interpretation of median (or any such calculation).
However, IIRC, Doug's point is that the numbers and calculations are
what they are and are technically valid. Considering what I understand
to be your criticism: I think the example you use above is a good
criticism of using the median. This is a case where a mean is more
useful. But in Doug's example, as you yourself identified, mean would
have been a bad idea. The median individual (or was it household?)
retirement savings amount gives a good idea on how people are doing in
general. In this case, the median would be misleading if it turns out
that 50% of America has median value + $1 saved away... here the median
may leave the false impression (though that is a human error in
misunderstanding the median) that a lot of people have a lot more than
the median value saved.
Your example above demonstrates how using the median can be misleading
(because of misunderstanding: note that the error in your example is not
mathematical. It is in the wrong calculation that median * no of sales =
total amount). A similar critique of the Fed numbers would be
informative. You have suggested that not the people who do the
calculations, but the people who present and use them are the one's
putting a spin on it. But Doug's point seems to be that what you have
available to you is (or so it seems) the direct results of the
calculations, not the spun version. Do you think that akin to the data
on climate change that was doctored/falsified/suppressed at higher
levels, the Fed info too might be tainted or partial?
--ravi
--
Support something better than yourself: ;-)
PeTA: http://www.peta.org/
GreenPeace: http://www.greenpeace.org/
If you have nothing better to do: http://platosbeard.org/
