Race and IQs

[Angrboda]

No, Anom has it right.  The null hypothesis is just the default counterpoint to any assertion/hypothesis being tested.  To prove an assertion/hypothesis you would have to show the null hypothesis to be incorrect.  

"Smoking causes cancer" ->  a claim.  The null hypothesis would be "Smoking does NOT cause cancer."  Through all the medical and scientific testing we've done, the null hypothesis was proven incorrect, and we know that smoking is a cause of cancer.

Using your example, isn't "There is a difference in IQ between races" a claim?  And the null is "There is no difference in IQ between races" ?

I think, if I'm understanding this, the failure to demonstrate a difference between the null hypothesis fails to support the claim, but does not on that account thus support the null hypothesis. It's a difference, if I understand things correctly, between accepting the null hypothesis ex hypothesi, and accepting it absolutely.