Thursday, June 2, 2011

The independence assumption: The difference between guilty and not guilty

The independence assumption is an important assumption for many statistical analyses.

Yes, a fair coin toss will give us 50-50 for heads or tails independent of what has happened before. Thus one coin toss is independent of another.

Two unrelated individuals should get independent scores on a computer game independent of one another. If the individuals are related then the independence assumption may be violated.

Here is a video written for a stats and methods unit but it has an important real life example in it where ignoring the independence assumption lead to a conviction of Sally Clark.




No comments:

Post a Comment