There is more to the TRCA data than only the p=0.48. Also the number of datapoints between the groups (before taking Increlex and after) is different as shown by Readi on the Yahoo forum. This is also a bad thing to do if you want to get good and comparable statistics.
Also concerning the p=0.48: a fifty-fifty chance that the groups don't correlate (which is what p actually means if I remember my statistics lessons correctly) is still quite insignificant.
Market Data 
Markets 
AI
