Tuesday, May 18, 2010

Cell Phone Brain Cancer and Statistics

A recent Globe and Mail article reported about a study called Interphone. It sampled 5,100 people, the supposed largest study of its kind, and claims that heavy use of cell phone may increase the risk of tumors. The point made by the study is that frequent cell phone users have a 40% higher chance of glioma, which is a type of tumor that starts in the brain or spine. Forty percent!!! That’s huge! Don’t you think that that size of increase would cause a noticeable number of people in your circle succumbing to this disease? Have you observed that to be true? After all, almost everyone and their kid has a cell phone.  To make the article scarier, it brings up the death of Edward Kennedy and asks the question whether his brain cancer was caused by excessive cell phone use.  
 
I’m sure you have heard of many similar studies with conflicting results. I am not a brain tumor expert and I’m not going to dispute the finding of this study but I want to put it into perspective as a statistics instructor. 

Statistical reports are created by researchers who make claims about the population using a analysis of a sample. For example, during elections campaigns, the public receives almost daily polls citing the  proportion of the voting population will vote for a particular political party. Do you realize that pollster base their numbers on a sample size of about 1000 people? Next time you see an election poll in print, look at the fine print to see the sample size. Statistically speaking, the larger the sample, the more sure the statisticians can be about their generalization with regard to the population. Due to lack of time, money or manpower, extremely large samples cannot be gathered but are always desired.

Getting back to the cell phone and brain cancer study, I want to alert you to another similar study.  You can read about this study at Scandinavian Study Finds No Cell Phone-Brain Cancer Link Over 9-Year Period. This study tracked 16 million people in Finland, Denmark, Norway, and Sweden over a people of 29 years. This group of researchers found no evidence that increased use of cell phones over a nine-year period led to more cases of brain cancer. Statistician or not, you know that a study using 16 million subjects would be a better than one that involved a sample size of 5,100. 

Why did the Interphone study claim that it was the largest study when there was one that was much larger? Maybe they were not aware of the other one? Maybe the two studies are so different from each other they cannot be compared. Maybe this is an example of selective journalism? Reporting on a study that does not scare people would not selling papers.

Here's what I want you to take away from this post:
  1. Question statistics you read. (How large was the sample? What was the study's limitations? What was the conditions of the study?
  2. Find a similar study for comparison purposes.
  3. Refer to multiple sources for new reports.
  You can read both studies for yourself and decide which one you believe.

No comments:

Post a Comment