Monday, November 2, 2009

Statistics - When is Diabetic News Believable?

Calling all folks who like to play bingo. Calling all folks who like to gamble now and then. Calling all people who believe everything they read in the news sources about new diabetes research as the gospel truth. If you do, you have just been scammed!

Jim, get serious. These folks want to believe. They have diabetes, don't like it and hang on every bit of positive news they can find. Okay, lets look at "odds" that play into research and believable results.

Take a quarter and flip it five times in a row. Its either going to be heads or tails each time. Everybody has done this at least once in their lives. You know, tonight you do the dishes. Heads I do them, tails you do them!

But, suppose you were going to Las Vegas for the weekend to enjoy a little "R&R." If you flipped that quarter five times and came up with five heads in a row, and you didn't know better, in Las Vegas you bet "heads" every-time in a would lose your shirt. The "Odds" in coin flipping is 50/50. The secret in odds, that you know, is the more trials(times you flipped the coin)the closer the results would be to 50/50.

In research, if you read an article where the experiment consisted of ten trials, the result has no validity. Its like getting five heads in coin flipping. Whatever that researcher says is nothing but an "opinion" and statistically has no validity.

In statistics, the number of trials is very important to support the results an investigator publishes. In order to state a 90% certainty for an experiment a very large number of trials must be run. Rarely does an investigator, seeking approval from the reading public, run a significant number of trials to validate what he is saying.

News leaked to the Press or TV, to gain exposure for a scientist, is not believable news unless backed up by statistics that are credible. Later, if other investigators back up the results of the research, you now have believable news.

Major publications, like The New England Journal of Medicine, have strict parameters before an article is accepted for publication. The source of the news is important too.

When is news believable news? Now you have a simplistic rule of thumb what research to put into a "circular file" and what research to become excited about.