Q20 - Quality control investigator: Upon testing

[steven.kantowitz]

To me, D and B seem to address very similar possible flaws. The only justification that I came up with for choosing D is that it is a more major issue.

Is there something obviously wrong about B that I am missing? Afterall, the arguement is presuming that the inspectors are just as likely to send defective as they are nondefective (and vice versa)...?

[bbirdwell]

The issue here is subtle.

The argument does not, in fact, presume (B). It's even quite possible that the argument presumes that the inspector is MORE likely to choose a NON-defective one. This would actually strengthen the argument's conclusion.

(D) however, is different in an important way, for if the inspectors purposely choose defective samples, then the % is definitely mis-representative, and therefore the conclusion has no leg to stand on, so to speak.

So, in abbreviated form, choice (B) says that the argument "assumes defective = non-defective," which isn't necessary.

(D) says the argument "assumes defective NOT MORE than non-defective," which is essential to the functioning of the argument.

Notice the difference in the language: "presumes" = "assumes" whereas "fails to consider" = "assumes not."

[noah]

Brian and I both started explaining this at the same moment, so I might as well add my explanation too so that I feel I'm useful in this world:

Tough one! The conclusion of this argument is that the manufacturer violated it's agreement to keep it's defect rate to below 5%.

Why? Because 20% of a sampling of the products were defective.

If you had to debate this, what would you say? I'd say "Well, who says your sampling was representative?!" Or, in more LSAT-like language, "You're assuming that the sampling was representative."

This is what (D) points out. Perhaps the inspectors tended to choose bad items, and thus the sample is not representative. Let's look at the tempting wrong answers!

(A) is a possible problem with sampling, but we have no reason to believe the sample was too small - we're told the sampling came from "various manufacturing locations." It's a big assumption to jump to saying that the sample size is too small.

(B) is super tempting! In essence it's saying that the argument assumes that the inspectors had a 50/50 chance of choosing a defective item and a 50/50 chance of choosing a nondefective item. The argument is not assuming even chances, it's assuming that the sample is representative.

Analogously, you were supposed to put only 80% burning hot red marbles and 20% cold blue marbles in a bag. John reached in his hand 100 times, and pulled out cold marbles 90 times. So, you didn't do what I asked!

What am I assuming? That John had a 50/50 chance of pulling a blue marble? No! I'm assuming that John's pulls were random. Perhaps he, for some strange reason, prefers cold marbles to burning hot ones.

(C) is quite tempting as well. However, we're told that the samples come from various sites. And, if (C) were true, we would then have to assume that the sampling somehow favored these sites - look at how much work I have to do to make (C) the right answer.

(E) is subtly irrelevant. We don't need the inspectors to have made the same number of visits - we need the sample to be representative!

[dtangie23]

I chose answer choice (C) for this one.

Suppose that only a few of the sites are responsible for the defective items. It's certainly a possibility that most of the samples collected could have been taken from these sites. That would mean that that 20 percent of the samples could have been defective, but the actual rate of defect (total) could have been lower than 5 percent. This is certainly something the argument is not considering.

It this choice incorrect because it is less probable to occur than field inspectors intentionally choosing flawed suspected defective products?

[noah]

It this choice incorrect because it is less probable to occur than field inspectors intentionally choosing flawed suspected defective products?

In a sense, yes. To make (C) work, you had to do a lot of "work":

It's certainly a possibility that most of the samples collected could have been taken from these sites.

And, you could just as easily have said that it's likely that the random sampling was truly random and representative, and so while there are some factories that make more defective items than others, their contribution of defective items is significant enough to compromise 20% of the total.

There's no way to "turn" (D) into a different sort of answer.

Make sense?

[jlhero]

i think one flaw is the argument overlooks the possibility that one of the factories can produce A LOT of products while other factories produce, say, 5% of the total.

so i think (C) explains this perfectly.

is (D) correct because LSAC assumes the relevant parties always have a conflict of interest?

so weird

[noah]

i think one flaw is the argument overlooks the possibility that one of the factories can produce A LOT of products while other factories produce, say, 5% of the total.

It's OK if the amount produced by different factories is skewed like that, as long as the sampling is proportional (that more of the samples comes from the factories that produce more). It's the sampling that has to be done "correctly."

Make sense?

[interestedintacos]

Here's why I eliminated B: because we don't know that the argument is assuming the field inspectors were just as likely to choose defective or nondefective. It's just not stated or even necessarily implied.

However, we can certainly say that the argument doesn't adequately address the one situation that would certainly ruin it: the sample overrepresenting defective products.

In short: we don't know the argument assumes they're perfectly representative, but we do know the argument doesn't adequately address the specific possibility of overrepresentation that would ruin it.

Also, not being perfectly representative would not kill the argument, as it could be the other way--it could be more nondefective products on average, which would strengthen the argument.

By the way, I do think "content" plays a small role here, as suggested by a previous poster. It helps to realize that an "inspector" is likely inspecting the objects in the field, not just sending back random objects. Why have an inspector do that--and why wouldn't the inspector be inspecting something? If the argument took out the inspectors and merely said 'objects were sent from various locations' it would be the same thing, but potentially a little harder to spot. In that case the equivalent of choice B would be: The argument assumes that the objects chosen are perfectly representative of the general production, and choice D would say: overlooks the possibility that the objects chosen overrepresented defectiveness. I think it would be even harder to pick the right answer, although D would of course still be correct (since it's the specific problem, and contained within B is also the possibility to strengthen the argument).

[interestedintacos]

I think that's where I went wrong, because my prephrase was "he's assuming that it was a true random sample

That pre-phrase matches choice B and is incorrect. He isn't assuming it's a perfect sample. He could also think the sample is underrepresentative of defects, meaning it provides even stronger support for his claim.

Is they key to (B) being wrong that the author is not assuming that the chances of selection were 50/50, but that the sample was representative of the actual proportion?

He DOES need to show the sample is not overrepresentative of defects; he does NOT need to show it is representative because he could show the sample is actually unrepresentative in the direction that HELPS HIM--that it underrepresents defectiveness, in which case the argument would be strengthened.

Again, strictly speaking, the author is not assuming the sample is representative, nor is he required to do that for the argument to hold. That's why B is incorrect. I think Mr. Birdwell put this well earlier.

[skapur777]

I'm gonna go against the grain here and say that I picked (A) which seems to be one of the more obviously wrong answer since everyone seemed to pick (B).

Perhaps I was inattentive but I thought I knew the flaw as soon as I finished the argument, who's to say that the samples that were defective are actually representative of ALL the items the supplier manufactures?

That's why I picked A. Even if 20 percent of the samples they tested in the laboratory, the actual number of samples could be extremellly small while the amount of products the supplier actually manufactures could be humongous and also non-defective.

So I thought A summed that up nicely. They based their conclusion on too small a sample of the items tested by the laboratory.

[noah]

I edited my initial explanation on the previous page to better address all the wrong answers. See if that clears it up. There's no reason to assume the sample is too small. It might have been, it might not have been.

We're not assuming the sample is a certain size, we're assuming it's representative. (D) directly addresses that assumption.

[americano1990]

Hi all

I just wanted to share with you guys the way I eliminated (B) and (C)-- supposedly the most tempting of all. I think some probably have touched on these ideas, but I would say them anyhow since they made things very clear cut.

Okay here are my thots

(B): it presumes something, and negation test is your best friend here. If you negate this you can arrive at two different possibilities

1. defective items were more likely to be chosen: in this case, the argument is weakened, because it indicates that the 20% mentioned in the stimulus is inflated. I.E it overestimates the error rate vis-a-vis the accepted rate of 5%.
BUT
hold on, as there is another possiblity

2. nondefective items were more likely to be chosen: this would mean that the 20% is an underestimate and that would support our claim that the supplier indeed is violating the contract.

Since negation of assumption should ALWAYS go against the argument, (B) cannot be one because it forks in two directions.


Why (C) is wrong: according to the argument, the supplier is the one that is overlooking the manu centers and this is why we are claiming the guilt of the supplier based on the performance of the manufacturing centers. THUS the distribution of errors among the centers is NONE OF OUR BUSINESS!

Lets say we want to hold Apple guilty for making bad iphones and we base off our claim on the defect percentage from apple manufactoring center from each of the 50 states taken AS A WHOLE. Would we care whether the defects were concentrated more in CALI or NEW YORK? NO! all we care about is the centers' performance as a whole.
With such an appeal to analogy, we can see why (C) is incorrect.

hope it helps

[vik]

The investigator is not assuming there is an equal chance of picking defective and non-defective parts. He is assuming a random sample. That means he is assuming a 5% chance of picking defective parts and a 95% chance of picking non-defective parts. Not 50/50, but 95/5.

[austindyoung]

The issue here is subtle.

The argument does not, in fact, presume (B). It's even quite possible that the argument presumes that the inspector is MORE likely to choose a NON-defective one. This would actually strengthen the argument's conclusion.

(D) however, is different in an important way, for if the inspectors purposely choose defective samples, then the % is definitely mis-representative, and therefore the conclusion has no leg to stand on, so to speak.

So, in abbreviated form, choice (B) says that the argument "assumes defective = non-defective," which isn't necessary.

(D) says the argument "assumes defective NOT MORE than non-defective," which is essential to the functioning of the argument.

Notice the difference in the language: "presumes" = "assumes" whereas "fails to consider" = "assumes not."

Sweet, so even if it says, "fails to consider" or "overlooks the possibility" you can just take that out, insert "assumes" and negate the rest of the answer choice so that it becomes like a necessary assumption question.

With that in mind (D) becomes much easier. But, if it has one of those phrases mentioned above in front of the answer choice- you can just remove it from the beginning of the answer- and wal-lah- the Negation Technique has already been done for you. So, if (D) is true, ya that does hurt the argument. Since the argument assumes representativeness, it is assuming that (D) did not happen- which is a flaw.

I spent way too much time on this problem....

[gplaya123]

I agree... I spent way too much time on this question too...

But I want my time not go wasted by teaching it to others so that their time don't go wasted

Let's identify the argument.

There is a supplier (let's call it... A). And it runs 3 factories (let's call it B C D). So let's imagine you are the CEO who has made a contract with A. Your employee or investigator brought 10 items from B C and D as a sample of items. Two were defected. Now you are pissed. Now, you are ready to sue the supplier because the defect rate supposed to be below 5% of ALL manufactured item. But according to sample, it's 20%. But your lawyer says "nope... you may be wrong."
How?

You may now understand the flaw due to capitalized words. You as a CEO assumed that the sample that the investigator has brought you is REPRESENTATIVE of ALL manufacturing items. However, could it be possible that your employee only picked out the ones that were defected? YES! If so, is it representive of ALL the manufactured items? NOPE!

Now let's go attack answer choices A to D (E is excluded because it's just so wrong).

A: All the right terms seem to be there but wrong because of one word: SMALL. We really don't know what is SMALL. Is 10 small like in my hypo? Could be. But may be not. Just because a sample has "small quantify" does not mean it's not representative. Remember, we are concerned representativeness not small quantity.

B: The most HATED ANSWER CHOICE EVAAAAAA. But here is the thing. Does the author assume that investigator was as likely to choose defected item as they were to choose non-defected item? HELL YEAH. Otherwise, how the hack would he assume that the sample is REPRESENTATIVE? But here is the thing. Even though B is true... it's not a FLAW. It really isn't. Imagine such assumption is true. Good. If it is true, as Brian has pointed out, it would straighten the argument to the max! If the investigator had 50% chance of picking bad item and 50% of good item and ended up 20% defect, that is a clear violation of contract. But... flaw is something that points out the weakness of an argument. Ironic much? Anyway, it's wrong because it helps the argument.

C: ANOTHER tempting choice. However, it works just like B. Think about it. Let's assume this answer choice is correct. So, it's saying that out of 3 factories B,C, and D, D were responsible for the 20% defects while B and C had 0% defect rate. Sounds pretty good doesn't it? But again, what's the contract? That overall defect rate should be 5%. According to sample, it's 20%! So no matter who is responsible for this defect, if the sample contains 20% defect, it's a violation. C is simply saying that "hey factor B and C are innocent." But who cares?

D) This is what we were looking for. Since, investigator was only picking out the defective ones, it's saying that the sample is biased.

Hope this helps.

[jones.mchandler]

The word "suspect" is tripping me up in D. The flaw is pretty obvious in this stimulus, but the ACs are all cleverly disguised.

In D it seems that you have to make the assumption that these inspectors are somehow nefarious maybe--that they want to choose items that they think are defective. So it seems that there's some issue with intention in D, which made me very uncomfortable with it as an AC.

I felt that C was closer at getting at the flaw bc the stimulus says "various manufacturing locations." So maybe they were sent to certain locations with faulty equipment that causes some sort of systematic flaw.

So what's up with the intention of these inspectors?

[ohthatpatrick]

I see your concern, since the psychology of the field inspectors feels new.

However, I don't quite agree that a field inspector has to be 'nefarious' to choose suspicious looking products. That's his/her job: to find items that DON'T meet quality control standards.

If you were a meat inspector, should you go about your job by just checking every 10th pile of meat, skipping over any rotting/discolored/fly-swarmed package of meat if it's not the next "random" sample?

No one ever said that these guys were "randomly" sampling, so it's perfectly reasonable that an inspector would slow down to more closely inspect the sketchiest looking items and then send THOSE out for testing.

Another thing to keep in mind is that when you see "overlooks the possibility" / "ignores the possibility" / "fails to consider" / "neglects to consider" start off a Flaw answer choice, that answer choice is essentially asking you a Weaken question.

I would process (C) and (D) simply as, "if true, does this weaken the argument?"

I would process (B) and (E) (presumes / takes for granted) simply as, "Is this a Necessary Assumption? Did the author have to assume this?"

So as you evaluate (D), you only have to ask yourself, "IF the field inspectors tend to choose items they think are defective, does that hurt the move from Prem to Conc?" Sure! Because the author was acting like the 20% defective rate came from a random, representative sample.

Hope this helps.

[jones.mchandler]

It definitely helps, and I appreciate the response. The info about how flaw ACs start off was particularly helpful.

Thank You!

[GodLovesUgly]

I'm having trouble understanding why A is so quickly eliminated. Here is my thinking...

20% of the samples tested proved defective

Supplier is contractually obligated to keep "defects" below 5%

Understanding this, could you argue that this is in fact a sampling issue? Maybe the "defects" were 20% on this shipment of product, but that doesn't mean that the supplier is negating his/her contractual obligation, rather although the % of "defects" was higher on this particular shipment, the supplier overall could still be limiting "defects" to under 5%

Ex. If I own a clothing store with similar stipulations from my supplier(s) and receive shipments every month, there could be 20%+ defective product in March and still over the course of a given year, the total % of defective product could be <5%??

Maybe I'm reading too much into this, but it seems like this could be a legitimate reason for identifying the flaw as having to do with sampling.

Thx

[ltownsjr]

I'm having trouble understanding why A is so quickly eliminated. Here is my thinking...

20% of the samples tested proved defective

Supplier is contractually obligated to keep "defects" below 5%

Understanding this, could you argue that this is in fact a sampling issue? Maybe the "defects" were 20% on this shipment of product, but that doesn't mean that the supplier is negating his/her contractual obligation, rather although the % of "defects" was higher on this particular shipment, the supplier overall could still be limiting "defects" to under 5%

Ex. If I own a clothing store with similar stipulations from my supplier(s) and receive shipments every month, there could be 20%+ defective product in March and still over the course of a given year, the total % of defective product could be <5%??

Maybe I'm reading too much into this, but it seems like this could be a legitimate reason for identifying the flaw as having to do with sampling.

Thx

Hi Godlovesugly, don't know if you're still around but basically what you said could be true from my perspective..but it could also be false. We need more information about the number of samples in order to say whether or not it's too small a sample size. Maybe that 20 percent equates to a total of 100 samples. Maybe that's a very representative sample. When you're doing a sample size you only need what could be considered a reasonable/representative set of the population you are trying to sample. What you said could be right, but what I said about the samples could be right as well. Remember our task for flaw questions is to expose the exact flaw that is made. Just because something could be the flaw, doesn't mean it is. Since there is not more information about the number of samples we simply can't tell either way. So how in the world can that be the flaw? It can't.

I hope this helps you, or someone else in the future!