Friday, August 7, 2009

Sample Sizes and Studies

Fascinating news if you're reading the Toronto Star. Apparently they've found that psychopaths have a 'pothole' in their brain activity. This the Star exclaims breathlessly, could lead to treatment of psychopaths or even early detection and crime prevention!

Wonderful news. If it was worth the paper it is written on.

What the Toronto Star doesn't tell you is that the sample size for this study was a whole 9 people. That's right, nine. You can literally count the number of sampled patients on your hands. Don't even need to take off your socks to count toes. But the Toronto Star doesn't tell you that, because then you might get suspicious about the results. Fortunately, other sources of news are far more responsible and do.

To explain my skepticism consider the humble penny. Ideally, when you flip the coin there is a 50% chance of the coin either ending up being heads or tails. Does that mean if you flip the coin twice you'll get one head and one tail? Of course not. You might get two heads or two tails.

The same logic applies to scientific studies.

Sure its possible that the studies results are valid and correct. However with a sample size of 9 it is impossible to know for sure. So all we have is another study that might show that there is a possible linkage between brain activity X and certain personalities.

Pardon me if I'm not impressed.

No comments:

Post a Comment