One of the FallaciousArgument''''''s, where an unstated individual or group is invoked as a (hearsay) witness to support the speaker's position.  Generally indicated with phrases like "some of us believe", or "lots of people think", or some other version of an existential quantifier.  (Either that or a denial of the universal quantifier--which is the contrapositive of the existential quantifier--i.e. "not everybody thinks so").

This sort of argument, like NoTrueScotsman, usually has a vacuous premise, which is almost always true--given any proposition, there is doubtless somebody who believes it true (or false), and the speaker can be construed to be part of the "some".  It is usually dubious whether or not it implies the conclusion.  Usually, there is no direct implication (as people thinking something to be true doesn't necessarily make it so), the intent of this argument is to imply that a consensus believes it to be so.  In many cases, the listener is intended to further infer that the AnonymousChoir holding such opinions are knowledgeable on the matter.

In short, it's a cheap form of ArgumentFromAuthority, where no specific experts (or pool of experts) are revealed.  A good way to tell the difference between ArgumentFromAuthority and AnonymousChoir is whether or not the claim of agreement can be falsified.  If it can (somebody quotes a particular expert, or makes a strong statement like "four out of five doctors agree"--claims which can be tested by asking the named expert or performing a survey), it's more likely ArgumentFromAuthority.  If the claim cannot be falsified--either because it's vacuous or ambiguous, then it's AnonymousChoir.

Another common use of this fallacy is to manufacture dissent to a proposition offered by one's opponent, so that the speaker may engage in ShiftingTheBurdenOfProof.  One example:

A:  <Foo> is an excellent technology for developing <Bletch>

B:  Not everybody thinks so; there are some people that think <Bar> is a better technology.  You'll have to prove to me that <Foo> is really better.

One way to deal with the AnonymousChoir is to try and out them--ask "who"?  Request specifics.  If the speaker refuses to (or is unable to) identify the AnonymousChoir with some specificity (or at least point to some evidence concerning the choir), or if the "choir" turns out to be a solo performer, then you can point out the error in the speaker's logic.

''The AnonymousChoir can to some extent also be used for the more legitimate purpose of asking for more evidence, though.  "You'll have to prove to me that <Foo> is really better" doesn't necessarily mean "I think you should believe that <Bar> is better"; it can simply mean "Your statements so far have not convinced me that <Foo> is better -- show me more of your data".  On the other hand, the latter phrasing could just be used instead.  Does someone have a specific opinion about that?''

* Your formulation of the statement--"You'll have to prove to me that <Foo> is really better", isn't AnonymousChoir, unless you add something like "many people disagree" or somesuch.  It may be ShiftingTheBurdenOfProof, but that's another fallacy.