Hi Koo. BDAZ gave the perfect definition.

A simple way I would put it is a p<0.05 means that the results of an experiment had less than a 1 in 20 probability of having happened by random chance (hence, inferentially it implies that the treatments applied to the test patients probably worked).

So a p=0.01 means that the experimental results had less than a 1 in 100 chance of occuring by random chance. This is 5 times stronger evidence than p=0.05 that the drug or treatment worked.

On the other side of it a p=0.2 means that there is a 1 in 5 chance (20%) that the results of the treatment were just random (i.e. very little support that the drug or treatment worked).

As BDAZ pointed out there really is no meaningful difference between p=0.055 and p=0.05 but the accepted standard is p = (or less than) 0.05.

Hope I didn't confuse things.

The levels are arbitrary and deals with accepted levels of risk.

Toinv