by ohthatpatrick Mon May 19, 2014 1:52 pm
Great questions!
I'm going to give you my best guess about the answers, but it should somewhat reassure you that it's possible to get any LSAT question right with even my somewhat impoverished understanding.
One thing to clarify at the outset is that correlation is not a black and white term. Mathematically, correlations are measured on a scale from -1 (negative correlation ... the more a, the less b) to 0 (no correlation ... a and b have no relationship) to 1 (positive correlation ... the more a, the more b).
But you can have correlations that are weaker than 1 or -1 but still significant.
Almost every correlation we see on LSAT is of this weaker form:
"Students who were chewing gum during the quiz were more likely than those who weren't to evaluate the quiz as easy."
What is an answer choice gave us this:
Harvey ranked the quiz as easy, and he was NOT chewing gum.
Would that weaken the correlation?
Yes, in a sense. But more importantly, it would not weaken the argument.
Harvey is a data point that pairs up easy + NOT chewing gum.
That means that we do NOT have a correlation of 1 for chewing gum and ranking easy.
But we may still have a correlation of .98 or something, which is still an incredibly strong correlation.
So for LSAT purposes, remember that when they phrase a correlation as
"Ppl who are A are more likely to be B"
or
"Ppl who are A tend to also be B"
there is still PLENTY of room for exceptions.
Because of this, you would never express a correlation with conditional logic. Correlations aren't guarantees; they're just trends.
As you were suggesting before, you can think of them as probabilistic statements.
If I know that Student X was chewing gum, I know that he/she probably evaluated the quiz as easier than did a student who didn't chew gum.
You can't really get into reversals/negations/contrapositives with correlations like these because you don't know the underlying numbers.
For instance, let's say
40 kids - gum .... 30 said easy, 10 said not easy
100 kids - no gum ... 50 said easy, 50 said not easy
This honors the original correlation we said:
Students who chewed gum were more likely to think the quiz was easy than students who didn't chew gum.
From this set of data, can I say:
"Students who thought the quiz was easy were more likely to have chewed gum"?
No! Of the 80 ppl who thought the test was easy, 50 were non-gum chewers.
Can I say:
"People who didn't chew gum were more likely to think the test was not easy"?
No! Of the 100 ppl who didn't chew gum, they were evenly split between easy and not-easy.
We could go on like this with other examples that SEEM like rephrasing of the same correlation but don't HAVE to be true.
That's why sometimes correct Strengthen answers will give you the "other side" of the correlation, something like
"Students who didn't chew gum were more likely to think the test was not easy".
This is a new fact! We didn't already know it from the original correlation, and now knowing it, we have even stronger evidence that there might be a connection between chewing gum and perceived ease of the test.
----
The way you were representing correlations was like this:
As A increases, B increases.
As stated, that sounds like a positive correlation with a value of 1.
Similarly, "As A increases, B decreases" sounds like a negative correlation of -1.
These are much stronger statements than the more common "ppl who are A tend to be B".
When we say "As A increases, B increases", picture an X-axis and a Y-axis and a line that slopes up and to the left.
You can totally read a graph like that in any direction.
If A goes up, B goes up.
If B goes up, A goes up.
If A goes down, B goes down.
If B goes down, A goes down.
Same thing for negative correlations, but obviously with A and B always having opposite effects.
But even though you could graph that statement and read in any direction, you really want to interpret it as conditional on LSAT.
Let's say we were told that
"as income level increases, happiness increases"
I could set up a graph with income level and happiness and see those two variables always rising or falling together.
But the way we really want to interpret that sort of statement on LSAT is like a conditional.
If more income --> then more happiness.
Essentially, more income is sufficient to bring about more happiness.
But maybe it's not required.
Maybe it is also true that
"as interpersonal love increases, happiness increases"
If you know someone's happiness has gone up, you can't be sure whether that person has more income, more love, or some other cause of happiness.
If you know someone's happiness has gone down, you can't be sure that the person lost income or lost love. Maybe something else impinged on the person's happiness while income and love stayed constant.
If someone's happiness went down, I can be sure that they did NOT increase their income and that they did NOT increase their love.
-----
Finally, if you have a causal chain of
A --> B --> C
you can definitely say that A indirectly causes C.
Depending on how an author's argument is worded we might care about that "indirect" or not.
If an author is claiming, "As you can see, sometimes A can bring about C" we'd probably have to agree.
If an author is claiming, "As you can see, A is what's responsible for C", we might be able to make some nuanced disagreement that A is only indirectly responsible for C.
But generally, yes, accept that A still causes C, even if by acting through some intermediary.
One example of that came to mind
PT37, S2, Q14
answer choice (C)
If A --> B --> C meant that only B is really causing C, then choice (C) would weaken and show us that
Eyestrain --> Headache
Instead we have
VDT --> eyestrain --> Headache
which actually strengthens the claim that VDT is causing the headache.
Hope this helps.
