I disagree with 1006a's rationale somewhat. Objectively speaking, although there was a drop in usage after 1970, Google ngrams shows that colored people is still used in the 21st century, and it is used almost as often as it was during the 20th century:
I restricted this to multi-word phrases predicated on the assumption that single words tend to be more popular than phrases in general, and generally more polysemous. Minorities is far more popular than any of these terms, and has applicability to mathematical contexts for instance.
By 2008, we can see that minority groups is almost on par with people of color, and people of color is far more popular than the closest equivalent, which is ethnic people, despite the fact that minority groups can include other classes of people. Moreover, unlike other racial terminology, the term's purpose was never inherently offensive. I do not think that use of the term in and of itself merits a flag, especially considering that it is much more popular than the closest equivalent, ethnic people, which almost flatlines. Also, considering how usage of black people and minority groups also dipped in a similar timeframe to colored people, the reason colored people is used less often today than it was in the past seems likely to be that we are discussing race less often than we were back then.
The closest thing I have found on our website to suggesting colored people is offensive is attributed to William Safire writing an unspecified October 28th 2001 article for the New York Times:
It's possible that the president coined the phrase; if so, it was on the analogy of women of color, a description adopted by many nonwhites. (Though colored people is dated and almost a slur, people of color is not in the least offensive.)
Suggesting that the term is dated seems almost contrary to my findings on ngrams, but more importantly there is a considerable difference between what is almost, and what is. It is a true statement, just as much as colored people isn't a slur. It does not presently show up on Wikipedia's List of Ethnic slurs, which makes sense since, insofar as I know, it was never meant as a slur.
Far more popular than the questin citing William Saphire is this answer citing George Carlin, who suggests that colored people and people of color should be considered equally offensive (or inoffensive), and prefers that we refer to africans in particular as black people.
What I take from all of this is that if we are going to appeal to contemporary sensibilities, you should use black people to refer specifically to africans, but colored people is probably acceptable in a broader sense. In my opinion, people of color and colored people are at worst, on the level of minced oaths, which makes sense considering that they were seemingly devised as euphemisms.
We're not talking about words that are as inherently offensive as the ones mentioned the last time a similar issue came up recently, and I don't think we should react anywhere near the same way.
What I Think Should Be Done in This Particular Case
I nevertheless suggest editing that example out of the post, despite my preliminary thoughts. Preferably after doing the poster the favor of replacing that example with a nonracial equivalent so that voting is not adversely impacted, or at least a comment explaining the edit. Minimally, a comment should be left.
Rationale for The Suggested Course of action
I do think that this is a minor transgression, likely made by accident and undeserving of any punitive action or answer deletion, but that it is nevertheless a transgression in need of editorial remedy in accordance to our policies. The quote should be removed, and preferably replaced with an equivalent context (pop song lyrics in this case) by the editor. I am less concerned about the language being used, and more concerned with how it is being used.
My overall interpretation of Stack Exchange's code of conduct is that we have very, very low tolerances for prospectively rude or demeaning behavior. Strictly speaking, the code of conduct has zero tolerance wording:
We don’t tolerate any language likely to offend or alienate people based on race, gender, sexual orientation, or religion—and those are just a few examples. When in doubt, just don’t.
However, sometimes analytical need overrides that in the academic community because our first responsibility is to accurately representing the truth as we understand it, whereas politeness is our secondary responsibility. I think it would be detrimental to our credibility have a zero tolerance policy and I believe Jeff Atwood afforded us some amount of leniency in recognition of this while he was in charge which we may interpret to be standing orders unless and until instructed otherwise, but in consideration of the zero tolerance language of the Code of Conduct, I do think we need strict compliance guidelines, similar to the ones we have for lack of attribution and misattribution: Edit to fix as a first resort if possible.
You may be wondering why I am referencing this policy if I do not think the word is likely to offend. Well, it is not so much what words are being used, and how the words are being used which I find to be problematic.
In Random House Webster's Dictionary, Second Edition there is a basic manual of style, with a subsection called Avoiding Insensitive and Offensive Language that Excludes or Unnecessarily Emphasizes Differences spanning from page 2223 to 2225, and this guidance given on page 2225 seems like the compromise we should use:
Reference to age, sex, religion, race, and the like should only be included if they are relevant.
The racial example given is:
Arab man denies assault charge[_________________]Man Avoids Assault Charge
In this example a racial epithet is being removed because it is irrelevant to the principle message, despite the fact that, at least in and of itself, Arab is not considered an offensive word in English, and especially not so when this was likely to have been written, between the years of 1989–2001 (initial and final copyright dates of my copy), meaning that was likely written in a pre-9/11 world. Given that this section is not part of the addenda, I think it is even probable that the provision was written in '89.
Now yes, an assault charge is different from saying "doo doo doo", but that merely affects the severity of the generalization, and I don't see why colored people would be used except to emphasize a distinction. Such generalizations are often considered offensive, even in nonracial contexts, and the fact that a racial generalization is being made transgresses these guidelines.
So in accordance to this guidance, we should weigh the prospective offensiveness of the term to the necessity of its use. First of all, in context, the necessity of the racial epithet, the manners of the time and the history of the word may render the quotation mostly, if not completely inoffensive, but that does not necessarily translate into the here and the now at Stack Exchange. We are not lyricists; we are analysts of the English Language.
In our capacity of analysts of the English Language, I think that it is necessary to accurately quote external sources in order to accurately analyze the language, but that the selection of quotations is also subject to the code of conduct, which would mean that the analytical necessity of the quotation the answer must also be considered when considering our code of conduct. If the quotation is not needful, then the use of the racial epithet in the post is not needful by proxy. Since our first responsibility is to analyze th English Language.
The question does not specifically regard that particular quotation, racial language in general or historical use, and the answer makes no effort to uniquely analyze that specific quotation to explain how the word go is being used on context or even how the usage pertains to personal experience with the language. As it stands, the answer makes no effort to analyze or explain the presence of any of those quotations at all in fact, and merely provides them as evidence of usage and axiomatic applicability. As a matter of fact, I have trouble seeing the utility of the answer, since it is simply General Reference, with a couple of seemingly random examples of usage added to disguise the fact. That makes them all interchangable with quotations from similar contexts of similar fame and quality, and means that the meaning of the answer is not impacted by the removal or replacement of the quotation.
If any effort to integrate the quotation into the answer was made, I would suggest leaving it in this case, especially since the choice of words is not especially offensive, but since no attempt to integrate it into commentary was made, I really see no reason for this answer to contain a racial generalization of any sort.
Only the quality of the answer would be adversely affected, since this would constitute the removal of evidence, and only if no replacement is furnished. In order to avoid this changing the results of voting, on an otherwise well received answer, the courteous thing for an editor to do would be to:
A. Explain why the edit is being made in the editing summary, so that a the answer isn't rolled back without due consideration.
B. Find an equivalent quotation to replace it so that the answer can be rated just the same as it would be prior to the edit.