oscar111
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Apr-27-06 11:54 PM
Original message |
End "ten to twenty percent" type of statistics: just say "fifteen percent" |
|
we are flooded with info from many fields of study.
Hard enough for us to recall just one number for a research finding.
dont fall for the latest fad of giving two numbers, when one will do fine.
Just give the average. Not some range.
|
petronius
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Apr-28-06 12:29 AM
Response to Original message |
1. But the range gives an approximation of the uncertainty in the statistic |
|
If people just give the average, we can't assess the precision of the stat (for example, 15 is the average of 10 to 20, but also of 0 to 30). I know what you're saying, but I'd prefer that people just stop spouting all the bullshit made up statistics that clog the info-ways; that would leave plenty of brainspace for necessary and meaningful data...
|
Whoa_Nelly
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Apr-28-06 03:03 AM
Response to Reply #1 |
|
Isn't that what it's all about? :rofl:
|
fishwax
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Apr-28-06 02:42 AM
Response to Original message |
2. but ranges can be useful |
|
and averages are not necessarily any more accurate, however appealing a single number may be ...
|
oscar111
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Apr-28-06 03:02 AM
Response to Reply #2 |
3. both of you are right, but we just dont have time for ranges |
|
outside of one's own field of expertise, are thousands of subfields of science and politics...
us laypeople... and we are ALL laypeople to those other experts...
just dont have time to even spend five seconds getting the feel of the range involved in every stat heard on tv.
As to accuracy, i did not mean that an average would have better precision,... it would not.. just that it would make it possible to recall the stat. Trying to recall two numbers is twice as hard.
Give us all a break.. and ensure your stat will be recallable by the harried millions working two jobs and raising a family.
|
realisticphish
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Apr-28-06 07:59 AM
Response to Original message |
|
:shrug: If you just said 15%, you would have to also give the 5% error. Scientific reports aren't going to change their precise standards just because people might find it hard to remember
|
ScreamingMeemie
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Apr-28-06 08:04 AM
Response to Original message |
6. In my job I have to use mean, median and modes...and range of values |
|
so there is some value in ranging things. We need them.
ex.
Your house is 20 years old. The house next door is a new construction. A house down the street is 70 years old, but the majority of homes on your street are in the 18-28 year age range. Mortgage officers need to know this when I turn in my report, because they may not want to lend money on the 70 year old house when there is evident spot lot building and a much lower range of ages as the norm. It will lower the remaining econcomic life of the home to be mortgaged. And they are not going to lend on a home with less than a 10 year remaining economic life span. Without ranges, I couldn't do my job. :)
|
DU
AdBot (1000+ posts) |
Thu May 09th 2024, 04:06 PM
Response to Original message |