Q22 - Columnist: Several recent studies

[dan]

22. (E)
Question type: Identify a Flaw
The flaw in this problem is primarily a mathematical one _ the author incorrectly associates a specific number of incidences with the general likelihood of an incident. To understand this mathematical flaw more specifically, imagine that the reports show that five times as many people were killed crossing with the light than were killed crossing against the light: if people cross with the light the same number of times people cross against the light, certainly we can say that crossing with the light is more dangerous. However, if people cross with the light 100 times for every time they cross against the light, the fact that the deaths are only five times greater would indicate that crossing with the light is much safer than crossing against it. (E) addresses the mathematical flaw in the argument, and is therefore the correct answer.

(A) is incorrect because it’s not clear that a bias exists, and it’s not clear how the bias would impact the findings.
(B) is incorrect because there is no causal relationship in the conclusion.
(C) is incorrect because it has no relation to the argument.
(D) is attractive _ however, a misrepresentative sample set is not the primary flaw in the original argument.

[ManhattanPrepLSAT1]

For those interested in a second explanation, this was originally posted May 4th, 2010.

mshermn wrote:

Fundamentally, the question feels very similar to one that mistakes a correlation for cause and effect relationship. Both answer choices (B) and (D) play off of causality. (B) states it exactly whereas (D) just talks about the effects being different at various locations.

The real problem with this argument has nothing to do with causality. The argument doesn't conclude that crossing with the light causes the danger. The conclusion is merely a statement of likelihood. So the problem is not with causation - the issue in this argument is mistaking an amount for a percentage. Just because more people die while crossing with the light than against the light, doesn't mean that the percentage of people who die while crossing with the light is greater than the percentage of people who die while crossing against the light.

Let's put some numbers on this.

Crossing with the Light
Total Crosses: 10,000
Deaths: 100
Likelihood of death: 1%

Crossing against the Light
Total Crosses: 50
Deaths: 1
Likelihood of death: 2%

More people die while crossing with the light than against it, but crossing against the light is still twice as dangerous as crossing with the light.

(A) is not accurate. There is no indication of potential bias.
(B) is the most tempting incorrect answer, but there is no causality posited in the conclusion.
(C) drives home the correlation/causation misperception but is not the issue with the argument.
(D) is not really a problem. The subject is not one city compared to another - this whole question was about North America. The statistics were about North America as was the conclusion.
(E) is the correct answer. If an equal number of people cross with the light, then the conclusion is true, but it's likely that a whole bunch more people cross with the light than against it.

[LSAT-Chang]

(B) is the most tempting incorrect answer, but there is no causality posited in the conclusion.

Hi Matt!
I picked (B) for this one, and I didn't even "think" about anything wrong with numbers in the argument.

When I originally read this, I thought "Okay so more people die when crossing with light than without (could just be coincidence or correlation) thus crossing against the light is less dangerous than crossing with the light (wow the author makes another correlation statement!)"

I thought this was a correlation/causal since the author is clearly concluding that something is more likely than something else just from the fact that there may or may not be a correlation. The "more pedestrians being killed with light" could just be a coincidence, but the author is basically saying that because "more" are killed from crossing with light, it should be the case that it is "less dangerous" to cross without light. So the author is assuming something like related with crossing with light and death and so it is in fact related, and that is why crossing without the light is less dangerous. Do I make sense?

I was honestly shocked to read the answer and explanation here since (E) was the second thing I eliminated after reading (A).. I can't believe I haven't even "thought of" and just "overlooked" the actual problem in the argument..

[hyewonkim89]

Hi Matt,

I was down to B and E and ended up picking E and got this question right.

But I still am a little confused why there is no causality in this passage.

To me, several studies showed the correlation between crossing with/against the light and danger, but the argument seemed to say crossing against the light won't cause much danger.

Please help!

Thanks always!

[ManhattanPrepLSAT1]

What words in the conclusion imply causation? Look for the usual suspects...

Causation
---------------
causes
because of
as a result of
leads to
has led to
contributes to
induces
stimulates
creates
produces
has the effect of
as a consequence of
due to
if you want to _______(effect), then you should _______ (cause)

[hyewonkim89]

I guess I just created a causality between crossing with/against the light and danger from a correlation... And was tempted by the answer choice B as you said in the previous post.

Your examples of the causation words help a lot.

Thanks!

[contropositive]

What words in the conclusion imply causation? Look for the usual suspects...

Causation
---------------
causes
because of
as a result of
leads to
has led to
contributes to
induces
stimulates
creates
produces
has the effect of
as a consequence of
due to
if you want to _______(effect), then you should _______ (cause)

i am still having a hard time seeing how this is not a causation. I understand why E is correct. I picked B because I saw an implied causation in the conclusion (crossing with the light is causing/having an impact on getting killed). I picked C during review, but now I see that B and C are both correct answers to a correlation/causation flaw which means they can't both be correct. Besides that reasoning, I still don't see how the conclusion is not a causation.
Aren't there two causation forms? 1) explicit causation and 2) implied causation. This one seemed like the implied causation.
For example, on the same section question 11, it felt like an implied causation (fish causing the likelihood of heart disease) which is why I picked D (the correct answer on that question) because it eliminated 3rd factor.

[maryadkins]

To say you're less in danger or more in danger under certain circumstances isn't actually a causal statement. It's stating a correlation, in other words.

For example, if I say, "You're more in danger of heart disease if you smoke than if you don't," that isn't implying causation. I realize it sounds like it is, and that's why this is an important nuance to understand on the LSAT. It's just saying there's a higher rate of HD correlated with smoking than there is with not smoking.

In this question, saying that crossing against the light is less dangerous isn't causal. It's just drawing a comparative correlation from the recent studies of numbers of people who die crossing the street.

Hope this helps!

[seychelles1718]

But isn't the INTERPRETATION of a PHENOMENON/CORRELATION typical form of a causal argument? As soon as I noticed the author concluded with an interpretation (less dangerous) of the correlation, I naturally thought this was a causal argument. Also, the author seems to think there are fewer deaths when crossing against the light is BECAUSE it's less dangerous to cross light against the light. So I thought the author was providing an explanation for a phenomenon. What did I miss??

[ohthatpatrick]

It DOES seem like some element of a causal flaw is present, but I think the real problem is that the evidence is not a correlation.

Raw numbers do not constitute correlations. LIKELIHOODS do.

If I say "there are more prisoners in Texas than in Delaware", that's not actually a correlation.

I'm not saying
"people who live in Texas are more likely to be prisoners than are people who live in Delaware"

You can assess the danger of something without assigning specific causality, even though that feels kinda like an oxymoron.

If I say 10% of people who travel to Russia get robbed while only 5% of people who travel to Spain get robbed, and then I conclude that it is therefore more dangerous to travel to Russia, then it certainly feels like I'm causally blaming some aspect of Russia.

After all, if I thought "People who are more likely to get robbed prefer traveling to Russia to traveling to Spain", then I wouldn't have said that "travel to Russia is more dangerous". I would have been chalking up the causality between the 10% vs. 5% to the travelers, not the destinations.

So I feel like I agree that the conclusion implies some level of causality.

This might be a rare case where (E) is a superior answer because the stem asks how the argument is MOST vulnerable to criticism.

Before we can even GET to the error that (B) would be describing, we have to first get past the confused thinking that (E) is describing.

IF WE GRANTED THE AUTHOR that a higher PERCENTAGE of people were killed crossing with the light, THEN we could argue what (B) is arguing in terms of how to explain that statistic.

But (E) is shutting down the argument farther upstream: dude, you haven't even convinced me yet that people are more likely to be killed crossing with the light.

[Heart Shaped Box]

It DOES seem like some element of a causal flaw is present, but I think the real problem is that the evidence is not a correlation.

Raw numbers do not constitute correlations. LIKELIHOODS do.

If I say "there are more prisoners in Texas than in Delaware", that's not actually a correlation.

I'm not saying
"people who live in Texas are more likely to be prisoners than are people who live in Delaware"

You can assess the danger of something without assigning specific causality, even though that feels kinda like an oxymoron.

If I say 10% of people who travel to Russia get robbed while only 5% of people who travel to Spain get robbed, and then I conclude that it is therefore more dangerous to travel to Russia, then it certainly feels like I'm causally blaming some aspect of Russia.

After all, if I thought "People who are more likely to get robbed prefer traveling to Russia to traveling to Spain", then I wouldn't have said that "travel to Russia is more dangerous". I would have been chalking up the causality between the 10% vs. 5% to the travelers, not the destinations.

So I feel like I agree that the conclusion implies some level of causality.

This might be a rare case where (E) is a superior answer because the stem asks how the argument is MOST vulnerable to criticism.

Before we can even GET to the error that (B) would be describing, we have to first get past the confused thinking that (E) is describing.

IF WE GRANTED THE AUTHOR that a higher PERCENTAGE of people were killed crossing with the light, THEN we could argue what (B) is arguing in terms of how to explain that statistic.

But (E) is shutting down the argument farther upstream: dude, you haven't even convinced me yet that people are more likely to be killed crossing with the light.

I agree with much of what you are saying Patrick and I don’t claim to be the expert to know casuation enough to the point I can definitively conclude whether or not or how much this Q has causality elements. However, I do feel like your analogy while good in spirit might be just a tiny bit off in analogizing the stimulus. I agree that there is no correlation implied in the numbber of prisoners in Texas vs that of Delaware, but that’s only bc the super set is humongous — there are hundreds if not thousands of prisons nationwide so simply one specific location (Texas) has a higher number does seem not imply a correlation. But the case in our stimulus is just a binary cut, that “either crossing with or against the lights”. There aren’t hundreds or thousands of different ways of crossing the light in which case your prisoner analogy would be more anaylogizing in my opinion. And if there are only “two” ways of doing a thing, and one of which has a higher number in a consequence, idk, it does seem suggest a weak correlation there. At least for me, when I read your analogy, i was like, oh yeah, totally agree, there is no correlation there. But when I read the stimulus, I get a different feeling. And after analyzing why, this is what I come up with. So personally, I do think there might be a slight correlation in the evidence but the conclusion isn’t committing a causation flaw. Instead, the conclusion is saying bc a higher “number” in something therefore a higher “likelihood” of risk/danger in that thing, which is in complete conformity with the rest of your analysis.