|
Authored by: Anonymous on Monday, August 20 2012 @ 07:33 AM EDT |
Only if the average is a median will there always be near enough 50% below
average (if at least one of the data set is the median, then it/them can't be
less, or more, than itself/themselves).
Using the data set {1, 1, 3, 4, 5, 7, 8, 9, 10, 992}
The [mode] average is 1 and most (80%) are above average
The [median] average is 6 and exactly 50% are above and 50% are below average
The [mean] average is 104 and most (90%) are below average.
Taking the average that most people assume by an unspecified average - the mean
average - it can be seen that 90% of the dataset is below average!
My dataset is not entirely unfounded (just taken to extremes to prove a point) -
if you take all the wages in a company from the part-time cleaner to the head of
the board of directors, you will find that most people are below average (mean
average, that is) in their pay as the few directors that have large pay packages
dramatically skew the distribution and so raise the mean significantly away from
the median.
If a set of data ever contains an unspecified average (ie it is not specified as
mean, median or mode) then be highly suspicious over what is trying to be proved
- it is highly probable that the mean and median averages will differ
significantly.[ Reply to This | Parent | # ]
|
|
|
|
|