This is a question that is, sadly, impossible to answer in an objective way. The problem is that it is very difficult to identify what statistics could tell us that it is working (or not) and it is difficult to determine how much affirmative action is contributing to those statistics.
First, we have to ask how we would measure affirmative action’s success or failure. Would we just measure how many non-whites are attending certain universities? Do we ask how many non-whites are in particular jobs? Do we need to look at something else like the overall average incomes of the groups that are supposed to benefit from affirmative action?
Second, even if we can agree on these measures, how do we determine how important affirmative action is in causing the results? For example, incomes for African Americans have risen in the decades since affirmative action programs have been implemented. Does affirmative action deserve the credit for this or is it due to things like decreasing discrimination or a general growth in the American economy?
Thus, we can only guess as to whether affirmative action is working. African Americans are clearly better off than they once were. However, we cannot know how much credit affirmative action should get. We also cannot tell how to balance that against the possibility that whites have become more antagonistic towards minorities because of these programs.