# RE: [IP] Statistical significance of kids with diabetes

```Gene Jorgenson [mailto:email @ redacted] wrote:

> small samples in statistics mean very llittle, especially when dealing
> with large populations.  The larger the sample taken, the
> more reliable
> the information.  Variation within sample populations can be
> very high,
> yet not be enough to be called "significantly different".

< Professional Mode = ON >

I'm afraid it's quite a bit more complicated than that.  Sufficient sample
size depends on many factors: the anticipated distribution of the data, the
relative frequency of the outcome being studied (prevalence of DM in this
case), the type of study being conducted, the proposed statistical model,
and many more factors.

Actually, the first question I ask scientists who come to me asking for help
in determining the necessary sample size is "How much money do you have?"
The second question is "How much will it cost to obtain one record of data?"
Divide the first by the second and you have the maximum sample size for the
study.  From there, we can determine the statistical power available, and

It is, by the way, quite possible to have too *big* a sample size.  If the
sample size is very large, then *any* difference, no matter how
insignificant clinically, will be statistically significant because the
statistical power will be virtually infinite.

> Depending on situations, statistics can be "bent" to show
> what you want,
> i.e. "Figures never lie, but liars figure."

The best read for a general understanding of statistics was written in 1955
and is still in press: _How To Lie With Statistics_ by Darrell Huff.  Paper
back is around \$5.00.  I'm pretty sure it's available on Amazon, which means
that it's available through the IP web site.  The ISBN is 0-393-09426-X.

Here's what some others have said about statistics:

"There are three kinds of lies: lies, damned lies, and statistics." -
Disraeli

"Statistical thinking will one day be as necessary for efficient citizenship
as the ability to read and write." - H. G. Wells

"It ain't so much the things we don't know that get us in trouble.  It's the
things we know that ain't so." - Artemus Ward

"Round numbers are always false." - Samuel Johnson

James Handsfield, PhD, MPH
Statistician
Division of Laboratory Systems
Public Health Practice Program Office
Centers for Disease Control & Prevention
email @ redacted / (770)488-8106
----------------------------------------------------------
for HELP or to subscribe/unsubscribe, contact: HELP@insulin-pumpers.org
send a DONATION http://www.Insulin-Pumpers.org/donate.shtml

```