#19 - Recent studies have demonstrated that smokers are

[mxro22]

Hi!

I understand why A is the right answer. However, I feel like C was also a potential correct answer.

C says "drinking caffeinated beverages is more strongly correlated with the development of heart disease than is smoking." In fact, strong correlation does not imply positive correlation; rather, it refers to the magnitude of the correlation. If two variables have correlation = -0.98, we would call that very strong correlation.

If caffeinated beverages were more strongly correlated with heart disease than was smoking, and the former pair's correlation was negative, the former correlation would overpower the latter and the argument would lose all force.

Thanks so much for all that this forum does! Interested to hear your thoughts on this.

[Christen Hammock]

Hey Cinnamon!

First, you're absolutely right that there's a relationship between correlation and causation, but it's a little more complicated than just conflating the two. If X always causes Y, and Y is only caused by X, then they would be perfectly correlated. If there are other variables involved, then that positive correlation would be weaker or nonexistent. That comes into play in this question. The conclusion states that there's a positive correlation between drinking caffeinated beverages and heart disease by effectively combining two other positive correlations (between smoking and heart disease, and between smoking and caffeinated beverages). That ignores the fact that these correlations could be hiding a more complex relationship—like Answer Choice (A), that among smokers, those who drink caffeinated beverages are less likely to develop heart disease.

In general, though, your understanding is right. So, to your other question, it's accurate to use the vocabulary of correlation when discussing that pattern of reduction in one variable being associated with reduction in another variable. The most important takeaway is the classic phrase: Correlation does not equal causation!

[Christen Hammock]

Hey Mxro!

Answer Choice (C) is actually an opposite answer! If drinking caffeinated beverages is more strongly correlated with heart disease than smoking, that supports the conclusion that there's a positive correlation between caffeine and heart disease.

[crispycrispr]

Hi,

I got this question right, but my thought process was very different than the explanation. I didn't go so far as to say that (A) makes sense and that it's possible caffeine mitigates the effects of heart disease--i feel like it's common sense that it does the exact opposite. However, I did see that the stimulus' weakness is in not mentioning any direct evidence between coffee and heart disease. So, I thought that (A) was right because it showed that smoking might be the main culprit, and that there may be no correlation between caffeine and drinking. I don't know if this thinking is wrong for the LSAT, though, so any feedback would be helpful, thanks!

[Adam Tyson]

That reasoning looks good to me, crispycrispr! Just because smoking correlates positively with two different things does not mean those things correlate positively with each other, and if we can show a negative correlation between those two things, or no correlation between them, the argument would be weakened. Answer A shows that negative correlation - boom!

[olenka.ballena@macaulay.cuny.edu]

Hi Powerscore,

I was stuck between A & B and ultimately incorrectly chose B, however I now understand better why A is the correct answer. However, in the original explanation, it was written that we are not trying to weaken a causal argument, as the author never said that drinking coffee causes heart disease (instead, it was a positive correlation). But if the author had made this a causal argument and did say that drinking coffee causes heart disease, would A still be the right answer, or would B still be correct?

I understood from this thread that because the author makes the relationship between coffee and heart disease a positive correlation, we can weaken this argument by proving that there's a negative correlation (what A does) or that caffeine has no effect on HD. However, if it was a causal argument (if the author said drinking coffee does cause HD), would A disprove that? My initial reasoning is yes, it still would because it undermines the argument that caffeine causes heart disease by showing that smokers who drink caffeinated beverages are less likely to develop heart disease. However, then I get caught up on the fact that A still says "less likely" which isn't 100% strong - meaning smokers who drink caffeine can still develop heart disease as a result.

Or would B correct if it was a causal argument? I feel like B might still be wrong because caffeine could still be a cause, but maybe there's a more important/stronger cause, which doesn't necessarily caffeine's role in heart development?

I apologize for any confusion; I'm just wondering what the difference between approaching the right answer is in causal versus correlation question for future questions. Is it perhaps the same approach, but just with different strengths? (i.e. causal arguments require a more define/absolute answer than correlation ones?)

Thanks in advance!

[Robert Carroll]

olenka.ballena,

It's difficult to even consider the situation because the stimulus is not a causal argument. The stimulus would have to be different to be one. How different? Whatever would be changed in the stimulus to make it causal, that would change what the stimulus is trying to do and thus change what would weaken the stimulus. So it's really difficult to talk about a counterfactual argument and what would or wouldn't weaken that.

With that in mind, I think answer choice (A) would probably weaken similar arguments that had causation in their conclusions, because a negative correlation is evidence against causation. Answer choice (B) is much more difficult to evaluate. That there may be more important factors doesn't exclude causation - X may be a more important cause of Y, but that doesn't mean that Z can't also be a cause. Thus, in our counterfactual situation, it would be important to know whether the conclusion thinks its cause is the primary cause or just any old cause. Again, since this isn't even happening in the real stimulus, there is no way to say what the conclusion of this counterfactual argument would have been, so I think that's all that can be said about it - it depends on what its conclusion is. What weakens an argument always depends on its conclusion.

Robert Carroll

[Davepix]

I circled back here because I thought no answer was worthy, read the stem again and started over wasted a few minutes and wanted to see if there was a better way... and was surprised to not see my thoughts mirrored.

In my translation:
"more likely" than equates to some, not most
caffeine <-s->smoking <-s-> heart disease
the flaw is that there is no correlation between caffeine and heart disease,

who knows what would happen if you did both, maybe there are no people that smoke and drink caffeine and have heart disease.

So I chose A upon enlightenment, correctly or by chance?

[Robert Carroll]

Davepix,

Your thought process is already embodied in previous posts on this thread. A correlation between X and Y and a correlation between X and Z does not make a correlation between Y and Z. That's not because "more likely" means "some", because it clearly means something more than that - we can truly say that some people who take the LSAT get a 180, but it's unclear what sense "there is a correlation between taking the LSAT and getting a 180" has, and it certainly doesn't equate to the quantification of "some". But as I said, your thought process was embodied in previous posts, insofar as it demonstrates that the correlation "more likely" is such a weak statement (and "some" is weak, so I get your thought process) that it doesn't really validate the conclusion. We could, I suppose, say that "some" and "more likely" and other relationships are too weak to validate connections like the following:

X bears some weak relation to Y
X bears some weak relation to Z

Therefore, Y bears that same weak relation to Z.

When the "weak relation" is "some" or "more likely" or a weak sort of correlation, arguments of that form are flawed.

Hope this helps!

Robert Carroll