What you are missing here is your unstated and unwarranted assumption that all of the 30 point discrepancy that you refer to is the result of irrational, racially motivated voting on the part of black voters.
That figure could, for example, be accounted for by 15% of black people preferentially voting for a candidate of their race, and 15% of white people doing exactly the same thing, in which case you would have no warrant for concluding that black people are "more racist". Or any other way of splitting the 30 points would also account for the figures. You, for some reason, have chosen to ascribe all 30 percentage points to bias on the part of blacks, and conclude from that that they are "more racist".
To put it another way, you have concluded that blacks are more racist based on the unproven premise that all of the racial bias apparent in this election can be ascribed to blacks. It needs a little more work to turn it into a formal petitio principii, but you're getting there.
This is actually a clever point you make (I haven't been able to say that in living memory
); that certain statistical biases can reinforce each other in the same apparent direction.
However I don't think your arithmetic is valid and you are not comparing comparable totals. You cannot take a percentage of one large group and simply add it to a percentage of a smaller group and say the result is meaningful.
If, in round numbers, the voting population of black Americans is about 10%, and at the moment we assign the candidates 50% each of the total Democratic voters (close enough), and assume an 80% bias on the part of black Americans, then Obama's 50% total can only be due to at most 8% of his total from blacks (assuming the 80% bias), and Hillary's at 2% from blacks.
Lets also say, for simplicity that there are just 100 total voters. 90 white and 10 black. If evenly split that would give them each 45 white and 5 black voters. That would mean Hillary has 47 white voters (let's ignore other groups for this argument, although they perhaps can contribute to the apparent effect) and Obama has 42.
That means that 30% of black voters "seem" to vote on race (3 out of the total of 10), while only 2 whites do (47 minus the theoretically equal 45). If there is a white bias (as there no doubt always is to some degree) it therefore can't be more than 2%.
If we take your 15% of white voters being biased then that would mean that Hillary has less than 35% of the unbiased white vote which is not something small enough to be missed by the pollsters and means that a large percentage of whites are blatantly lying when they claim to be unbiased by color (in the upper 90% if my memory serves me right in a recent poll, which fits with the above 2% estimate worked backwards).
As to your final point regarding motivation and racism, there are degrees and categories of racism, some malicious and deliberate, some unconscious or cultural, but in the final analysis I am still looking to be convinced that this apparent discrepancy doesn't suggest that black are more likely to vote just for their own color than whites.
