Official AP Statistics Thread 2012

<p>I just checked out that link from Shroud and did all the problems.</p>

<p>Got #17 and #18 wrong, anyone know how to do #17 specifically?</p>

<p>Good luck everyone:)</p>

<p>Any predictions for FRQ’s tomorrow?</p>

<p>Solution to #17 goes like this (I think):</p>

<p>200=1.86(7200/sqrt(n))
n=4483.6416.</p>

<p>The question asks for the least value of n that makes the margin of error less than 200. A number smaller than 4483 would make n larger than 200, so the next largest choice over 4483 is the correct answer. In this case the answer is 5200, C.</p>

<p>As for tomorrow, I’m guessing that there will be a linear regression problem, a creating/using a normal/binomial model, some hypothesis testing, but I doubt they’d ask for chi squares tests on the FRQ. Of course, I know nothing. Just based on old FRQs</p>

<p>Wouldn’t it be 1.96 for the 95% value instead of 1.86?</p>

<p>Yo guys, I have a question. In the Barrons book, whenever they use random number tables for simulations, the tables are REALLY big. Are they supposed to be that big for the actual test? If not, how big? And how would I do it on my calculator (just use the random # generator?)</p>

<p>I think I may be confusing myself, because the Barron’s didn’t make it clear. Are the tables provided to you or do you make them yourself?</p>

<p>@Monepo The tables are provided with your formula sheet which you have access to for the entire test.</p>

<p>Ok thanks! And should we use the whole table in our simulation?</p>

<p>No, for example if the probability of scoring a three throw is .7 then you would make the numbers 1-7(inclusive) stand for making the throw and the numbers 8-10(inclusive) stand for missing the throw. You do this following standard simulation protocol until you have ran a certain number of simulations which should be stated in the problem. So you really should only be using as much of the table as you need, which I have never seen exceed 2 lines of 15 digits for a problem.</p>

<p>Ok, so they’ll tell us how many trials to run. Thanks guys! I know I sound like a total noob right now haha.</p>

<p>Quick question. In a linear regression test if r^2=.52 would you say the predictions will be reliable?</p>

<p>Ive heard that the barron’s practice test is harder than the actual test, is this true? I’ve been doing pretty terrible on the barrons multiple choice, like getting only 20-21 on the multiple choice. Does that put in me in line for a 3/4? Also this may seem like a dumb question, but can you tell what the mean is around on a boxplot?</p>

<p>@zachb757
I think that would not be very reliable.</p>

<p>@ nervous jim do you know approximately the minimum cut-off for r^2 in order for a line to be considered reliable?</p>

<p>Well, I think we should look at the r value right? I thought that r^2 is tells you how much variation of the y is accounted for by the x. If thats true, then a correlation(r) greater than .8 is generally considered good i think.</p>

<p>Thanks, but I was taught the coefficient of determination determined the reliability of a line. So I’ll interpret your answer as approximately .64.</p>

<p>In the free response, when do we write the 4 step process of defining hypothesis, checking conditions, correct mechanics, and conclusion. I don’t when to write that. Can someone explain please?</p>

<p>@zachb757
Don’t quote me on that though, I could be wrong. My teacher didn’t really teach me what coefficent of determination was , he just said it was how much variation of the y was represented in the x variable.</p>

<p>@nervousjim Barron’s practice tests tend to be harder than the real thing, but if you get 20 MC right and do approximately as well on the FRQS then you can definitely get a high 3 or a low 4. However, some people find that they perform better under the pressure of a real exam or that they perform worse under the pressure.</p>