Q6 - In the summer of 1936

[interestedintacos]

I find it difficult to choose between answer choice C (the credited response) and D. We're asked to resolve the paradox that the poll showed Landon would win big, but Roosevelt ended up winning big.

I understand that C highlights something that MIGHT resolve the paradox, but it could just as easily not resolve it at all. It could be the case that telephone owning voters voted for Roosevelt at the same margin as non-telephone owning voters.

D is similar to me. It's possible that poll's glaring omission of political affiliations meant nothing--that in the end, for instance, the poll actually got an equal number of Democrats and Republicans, or that the vote for Roosevelt wasn't mostly based on party lines.

But which one is more likely to be a major factor in the polls inaccuracy?! Political parties! The standard assumption would be political parties play a more important role in the way people end up voting than telephone ownership!

How can the telephone ownership issue more likely resolve the paradox?

[interestedintacos]

After reviewing it again I'd say the reason is that D would require an extra step to know we have an unrepresentative survey, while C doesn't require an extra step. Is that right?

Even if there is no difference in the end between telephone owning voters and non-telephone owning voters in their support for Landon/Roosevelt, this point at least provides the best hope of resolving the paradox because we know in one area for sure that the poll wasn't representative.

With D we don't know for sure that the poll isn't representative in terms of political parties--we would need an extra piece of info--like that in fact many more Democrats were polled than Republicans. Also I suppose since the poll included a wide variety of respondents, rural urban, m/f, many states, the presumption would be it would likely include relative numbers of each party, while again, we know for sure that it doesn't include AT ALL non-telephone owning voters.

So I suppose I see it now. I'm still not sure how to learn from this in future 'resolve the paradox' questions though. A lot of these questions seem pretty slippery and open the door for more dubious reasoning.

[ManhattanPrepLSAT1]

Also I suppose since the poll included a wide variety of respondents, rural urban, m/f, many states, the presumption would be it would likely include relative numbers of each party

That's a better way of looking at this one. We know that one way to explain the inaccurate prediction is to show that the survey was unrepresentative. Answer choice (C) calls into question the representativeness of the survey because now poorer people are excluded from the survey, yet they can still vote. Answer choice (D) does not necessarily imply that the survey was not representative. You're correct about the additional assumption required by answer choice (D). Only if we discovered that the political affiliations of the respondents were split in favor of the party represented by Alfred Landon, that people vote according to their political affiliations, and that the political affiliations of the public at large were actually split in favor of Franklin Roosevelt could we use the information in answer choice (D) to explain the inaccurate prediction.

There's just too much assuming going on in answer choice (D), while answer choice (C) doesn't require a series of secondary assumptions. That makes answer choice (C) the correct answer. Another thing one might notice is that answer choice (C) plays off of a story that is commonly told about this event. Often, the correct answer to these early questions, comes from information commonly related to the event.

Let me know if you still have another question on this one!

[interestedintacos]

I think honestly it must be my background assumptions from all my work in politics that undercut me here.

Apparently this story pops up on another test (or somewhere in LSAT prep materials), and it turns out that the core reason the survey was not representative (I'm talking about the real history now) was because of the political party issue. The fact that telephone owners had been polled meant that a much higher number of Republicans were polled than Democrats--Republicans having a much higher percentage of the more well off voters who owned telephones.

The fundamental flaw behind the survey was that it did not not accounting for party affiliation, which anyone can pick up on--telephone ownership is just a clue in to party affiliation. Nothing is more important in political polling than political party, and any poll that doesn't include it is clearly going to be flawed.

[karen_chu22]

Why is Choice B wrong?

[etwcho]

Choice B is wrong because the survey is correct to be representative of people who are eligible to vote. The survey would misrepresent the voting population if it asked infants and young teens who they planned to vote for, because they simply can't vote.

Similarly, if the goal of the survey is to figure out sales forecast for particular golf clubs, then the survey should represent people who actually play golf.

[sportsfan8491]

I know that the distinction between why answer choice (C) was correct and why answer choice (D) was incorrect was already made by mattsherman, but I'd like to share my thoughts on this one because I think it might be helpful to some. One thing I noticed that allowed me to rule out answer choice (D) 'uberfast' is that this answer choice seemed like it was trying to introduce the concept of biased questioning, without actually doing so.

The individuals conducting surveys are free to ask whatever questions they want to ask, so long as the questions aren't biased in any fashion (i.e. the interviewers aren't asking "leading questions"). If it can be shown that the interviewers were asking biased questions or were purposely excluding questions to the detriment of the survey's accuracy, then I think we'd have reason to believe that the survey was unrepresentative.

However, the fact that political affiliation wasn't asked about doesn't introduce any notion of biased questioning because, for some individuals, political affiliation might not have any impact on the way they vote, especially if they dislike one political candidate due to his/her specific policies (excluding this particular question doesn't seem detrimental to a survey that's specifically asking about which candidate someone will vote for in an upcoming election).

[mjacob0511]

A varied telephone poll predicted AL would win but FR won in a landslide. We are looking to explain this apparently conflicting set of events.

(A) So what? Who cares about the interviewers?
(B) So basically trying to cry unrepresentative because it only polled those who would be eligible to vote, but I think it’s backward. That is exactly what makes it a good sample; it only took into account people who would be eligible. Why should it take into account people who won’t be eligible at the time of the voting? This does not explain the discrepancy.
(C) Aha so it really wasn’t a representative sample! It only took into account people who could afford cellphones. Maybe that meant only rich people and they were a minority. Perfect.
(D) But they said who they would vote for!
(E) Reasons are irrelevant.