That's a common misinterpretation of the results of statistical tests. You can never conclude from that type of test that two things are not correlated, you can only conclude that the data you have does not suggest a correlation ("fail to reject the null hypothesis").