Ocala.com Bloggers
- Broken News
- Bill Thompson
- Digital Editor
- Eric Barnes
- Frame 37
- Ocala.com Multimedia
- Marion Politics
- Newsroom
- Observations
- Naseem S. Miller
- OcalaDay
- Joe Byrnes
- Read My E-Mail
- Allen Parson
- Running Wide Open
- Joe Vanhoose
- Speaking of Business...
- Dr. Philip R. Geist
- The Bowling Blog
- Debbie Whitten
- The Green Zone
- Dave Rhea
- The Sports Blog
- Gregory Broome
- What is that?
- Newsroom
 |
Search Blogs
|
 |
|
|
|
|
Blog Archives
|
|
|
|
|
|
|
So, is statistical science leading us down the garden path?
|
|
|
|
|
|
|
Location: Blogs Sly Comments |
 |
| Posted by: Emory Schley |
5/28/2007 10:50 PM |
I’m not much of a mathematician, and that subject has always been a shaky one for me. And a quick glimpse into any reputable text on statistical science will show you enough math to make your head spin completely off your shoulders and screw itself at least half-way into the ground.
So, I’m not exactly an authority on statistics and the methods used to compile them, either. I will say, though, that while my confidence level in higher mathematics is pretty solid, my confidence in statistics is down toward the bottom of the scale, at best. I suppose the reason why is that I’ve seen many so-called studies that didn’t seem to make much sense to me. That situation has remained unchanged, largely because they keep rolling out study after study that doesn't quite seem to measure up to reality. I know that a supposed key to rock-solid statistical studies is the quality of the sample, or put another way, carefully choosing whatever people you want to take part in a particular study.
The biggest problem, as I see it, is that it would seem to me to be almost impossible to ascertain exactly what it is you’re measuring. Take, for example, a situation like this: You round up 100 people, being careful to ensure there are 50 males and 50 females. Then you ask each of them if they like ice cream. Let’s say 68 say yes, 30 say no, and 2 say they don’t know.
Now, you send a poll-taker throughout their ranks writing down each individual’s eye color. So, let’s say that 50 have brown eyes, 25 have blue eyes, and 25 have green eyes. Now you map each individual’s preference for ice cream to their eye color. Then you might be able to say that based on your research, 45 percent of people with brown eyes prefer ice cream, while only 10 percent of those with blue or green eyes care for the treat.
So, have you proven your premise? I think not, and I think not because if you changed the group, the statistics would most likely swing in another direction. I can see where you can be quite specific (and correct) in your premise AS it applies to that one small group, but trying to extend those same conclusions to a far greater population seems, to me at least, to be pretty preposterous.
But yet, isn’t that exactly what they do in coming up with TV ratings? They poll 1,000 households and try to make those figures represent the habits or practices of several hundred million people.
Hmmm... doesn’t seem very plausible to me. How about you? |
|
| Permalink |
Trackback |
|
|
|
|