1 |
On Wed, Oct 27, 2004 at 07:15:04PM -0400 or thereabouts, John Davis wrote: |
2 |
> I strongly abject to picking an arbitrary cut off though (33%). I would |
3 |
> rather that we use some sort of statistical analysis (simple central |
4 |
> tendency measures will work). The process is much more scientific, and |
5 |
> it ensures that our sampling population is represented fairly. |
6 |
|
7 |
I don't care if we pick an arbitrary number or fire up a copy of matlab and |
8 |
go to town. (though personally, I think we need to be careful about making |
9 |
this more difficult or involved than it needs to be.) The two important |
10 |
points in my mind are: |
11 |
|
12 |
* The criteria that we're going to use need to be stated up front, as part |
13 |
of the GLEP |
14 |
* The criteria need to be very clear-cut so there's no arguing down the |
15 |
road about whether the service should continue. |
16 |
|
17 |
Overall, I think the GLEP is solid and it's something that I personally |
18 |
support, as long as we can include objective, measurable criteria for |
19 |
success. |
20 |
|
21 |
--kurt |