Calculating averages but without the extreme/exceptional data points

[davy]

Dear All,

I need to calculate the average time for a deliveries but without the
exceptional data points. For an example ,typically I will receive my
order in 7 days, but if I pay for express delivery costs , I receive
my order in 2 days. If there is a delay in the postal service or my
supplier has no stock then delivery could be in 14 days.

My question, is how do I calculate the average delivery time (via
excel) but remove the early deliveries. As I deliberately requested a
short delivery time and these deliveries seem to weight the average
delivery time unfairly. (The data set is very small, typically less
that 10 data points).

regards

Davy

 

[Toppers]

=AVERAGE(IF(A1:A20>2,A1:A20))

Enter with Ctrl+Shift+Enter

 

[Bruce Sinclair]

Dear All,

I need to calculate the average time for a deliveries but without the
exceptional data points. For an example ,typically I will receive my
order in 7 days, but if I pay for express delivery costs , I receive
my order in 2 days. If there is a delay in the postal service or my
supplier has no stock then delivery could be in 14 days.

My question, is how do I calculate the average delivery time (via
excel) but remove the early deliveries. As I deliberately requested a
short delivery time and these deliveries seem to weight the average
delivery time unfairly. (The data set is very small, typically less
that 10 data points).

Well ... given the small amount of data, the easiest way would probably be
to leave the points you don't want to include out manually. Note that this
is not good stats practice in general.

If you want to get a 'feel' for how much difference the big and small
numbers have on your average (or for ease of use if you had more data), you
could set up a helper colum to help you sort the ones you want then include
and/or exclude at your lease ... but remember .. .statisitcians will hit
you if you tell them this is what you are doing .

You might want to look at medians. If you want an idea of how long it
"usually takes" this might give you a better guess. It often does if the
data is bounded at one end and not at the other (eg salary/wages and
delivery times).

HTH