Recording executive: Many musicians resent Web sites that allow peopl

­Understanding the argument - 
­Recording executive: Many musicians resent Web sites that allow people to copy music free of charge. - Fact/Background Info.  
The musicians argue, among other things, that each person who chooses to copy a song from such a Web site represents less profit for the song's creator. - Musicians claim.
However, by providing free publicity for the musician, the widespread copying of a song over the Internet appears to increase record sales. - Conclusion. Widespread copying causes an increase in record sales. 
There is a strong correlation between increases in the popularity of a song on music-copying Web sites and increases in sales of the album containing that song. - Supporting premises. 

Basically, the argument is that A and B are correlated. A causes B. 

Option Elimination - Flaw

A. It overlooks the possibility that two correlated phenomena stem independently from some third factor and that neither causes the other. - Exactly. A and B are correlated, but that doesn't mean A causes B. OK. 

B. It fails to adequately address the possibility that even if a phenomenon causes a certain effect in some instances, it may have the opposite effect in other instances. - "It may have the opposite effect in other instances" is out of scope. We are talking about A causes B. 

C. It takes for granted that if a practice increases overall record sales, then musicians have no legitimate reason to resent that practice. - Out of scope. 

D. It requires the assumption, for which no support is provided, that significant sales of an album containing a song seldom occur before that song becomes popular on music-copying Web sites. - It says A occurs before B. Yes, this assumption or the missing premise strengthens the conclusion. But our job here is not to find an assumption or the missing premise. Our job is to find the flaw. The flaw is as A and B are correlated, A causes B. Classic assumed causality flaw. No wonder this is a popular wrong option. 

E. It overlooks the possibility that most of the songs in albums that are purchased have not previously been copied by the purchasers from any music-copying Web site. - The core of this option is that "most of the songs in albums have not previously been copied." It weakens the premise given in support of the conclusion. Basically, it says >=51% of the purchased songs have not been copied, weakening the correlation, which is a fact. On GMAT, facts are respected and not challenged. Distortion. 

gmatophobia - I didn't get why Option D is wrong. What if significant sales (indirectly gain in popularity) is what is causing more people to copy music free of charge. Doesn't this reverse the causation that "widespread copying of a song over the Internet appears to increase record sales".

I think choice D is not a flaw because requiring an assumption is not a "flaw" - it is part of the reasoning. If this choice would say that the author overlooks the possibility that increases in music sales cause an increase in music's popularity, then this choice would be correct.

Can someone explain why E is not the correct answer.

If the increase in sales of albums is because of other songs which have not been copied previously and not due to the songs which have been copied then this option clearly weakens the argument. It is showing that there is another reason for the increase in record sales.

Choice E talks about albums that are “purchased”. This is not a proper distinction as it is a very generalized claim. The argument talks about albums whose “sales are high” not every album that’s ever purchased.
So it’s out of scope choice

Posted from my mobile device

­Hi experts Bunuel MartyTargetTestPrep KarishmaB
Please explain this question, especially the correct reason to eliminate option D.­

This argument relies on a classic flaw: correlation vs. causation. The author uses a correlation between X and Y (increase in copying, increase in legit sales) to attempt to prove that X causes Y (increase in copying CAUSES increase in legit sales). Any time we see this flaw, we can identify a few alternatives. Maybe X doesn't cause Y, but rather one of the following alternatives is true:

*Y causes X. (In other words, maybe increased sales lead to increased copying, not the other way around).
*Z causes X and Y. (In other words, some other factor, such as publicity, increases both copying and legit sales, and neither causes the other.)
*Coincidence. (These just happened to correlate for no reason.) This usually only applies when the data set is small. For instance, if I turn the lights off and the phone rings, it may not be that one thing caused the other--it may just be a coincidence. But if my phone rings EVERY time the lights go out, it's hard to write it off as a coincidence.

One other thing before we get to the answers. This is a FLAW question, a type we don't see too frequently on the GMAT. What many people (including some folks on this thread) don't realize about flaw questions is that their answers can come in three types. 1) They can point out an assumption, using terms such as "assumes," "presumes," "takes for granted," or "fails to establish." C and D are both assumption-type answers. 2) They can point out potential weakens, using terms such as "overlooks the possibility," "neglects," or "fails to address." A, B, and E are all weaken-type answers. 3) They can describe the flaw in abstract terms, using descriptions such as "conflates correlation and causation" or "draws a general conclusion based on an a sample that is unlikely to be representative." Although A-B are written directly as weakens, they also feature abstract descriptions, with no mention of the actual terms from the argument, so they fall a bit into this category, too.

A) CORRECT. There's the "Z causes X and Y" that we predicted. This is phrased as a weaken, and if a third factor caused both of the increases, then it would not be clear that one caused the other. This makes A a successful weaken, so it correctly describes a flaw in the reasoning.

B) First, this ignores the fact that the premise is about song downloads in general, not just a few isolated cases. They may well have considered that a different result is possible, but found that this doesn't happen or rarely happens, so it's not something that the argument clearly overlooked. Second, the conclusion only says that copying "appears to increase" record sales, not that it must always do so and could never have the opposite result. In any case, since the argument never proves causation in the first place, it doesn't seem pressing for us to point out that causation may not ALWAYS happen. It may be that it doesn't happen at all!

C) This is addressing a further conclusion that the author never makes. The author mentions the resentment that musicians feel, but only argues with the part about sales. The author never concludes that musicians have no reason to resent free copying.

D) This seems pretty closely related to our "Y causes X" alternative. Since it's phrased as an assumption, we can test it by negation: is it a problem for the argument if significant sales of an album OFTEN (rather than seldom) occur before the songs become popular on copying sites? Well . . . no. The key difference here is a subtle one, but the correlation we're given is about INCREASES, not "significant sales." So even if there are significant sales (let's call that 100k in downloads, just to have a figure) before the copying takes off, it could still be that an INCREASE in copying (from near-zero to, say, 10k per day) could still lead to an INCREASE in legit sales. Also, D tells us nothing about the nearness of these sales to the increase. If the copying happens, say, a full year after the strong initial sales, then the correlation in the premise is not going to be at stake. According to the premise, we'd still expect to see an increase in sales near the time of the increase in copying.

E) This seems to be trying to attack the connection between copying and legit sales, but the author never said that the SAME people will get the songs both ways. (Why would they?) We just need to know whether more copying leads to more sales, and E doesn't give us any further information to help figure that out.