<p>Could someone explain to me the difference between an influetial point and an outlier in terms of linear regression? My teacher explained it rather vaguely in class and I’m not certain how to graphically distinguish one from another. Thanks in advance!</p>
<p>^
Influencial point: A point that greatly contributes to the slope of the least squares regression line… if it were taken away the slope would change greatly.</p>
<p>Outlier: Any data point not near all the others in a regression model. It can throw off the slope of the LSRL. You eyeball outliers… there is not formula to find them.</p>
<p>I guess you could say “outliers are influencial.”</p>
<p>Thanks for the prompt reply, Cody. The review book I own made me doubt my previous understanding, I suppose, because what you’ve just said is what I assumed from class. Again, thanks for helping this newbie out.</p>
<p>^ No worries. Good luck tomorrow!!</p>
<p>You too. 22.5 hours till test start and still cramming on my part. I’ll probably pop in again if I need more clarifications on some older material.</p>
<p>22.5 hours till your test starts? What? It’s at noon tomorrow…</p>
<p>Ah, my apologies, I meant 10.5 hours-- very stupid arithmetic mistake. : )</p>
<p>anyone know how to do question 14 from ap stats mc/??
<a href=“Supporting Students from Day One to Exam Day – AP Central | College Board”>Supporting Students from Day One to Exam Day – AP Central | College Board;
like page 25 question 14
plz help me i don’t how to solve it :(((</p>
<p>^ What do means and standard deviation have to do with lines of regression? I know that question’s possible, but I’m still inclined to say that it’s not. You need a slope…</p>
<p>Well, first, you know that the intercept is given by ybar - (slope) * xbar. (It’s on your formula sheet if you don’t.) So you just have to go through and check that.</p>
<p>When you do that, you’ll find that both a and d satisfy that equation. So how do you know which one is right? Remember also that the slope is given by r * sy/sx. Now, you don’t know what r IS, but you know it has to be between 1 and -1, so your slope has to be between sy/sx and -sy/sx. So that leaves you with just answer d.</p>
<p>hey thx for ur reply
but i still dont get it
where’s that formula
is it b0=ybar -b1xbar ??
y bar in the problem is 10 right?
what about the slope
can u input value in ur explanation plz
thanks very much</p>
<p>^ This is probably too late already, but that formula is the 6th one on the first page of the formula sheet.</p>
<p>The slope or b1 can be calculated using the formula r x sy/sx. As amarkov said, you don’t know what r is, but you do know that it has to be in between 1 and -1. So b1 has to be in between -1 x 10/4 = -2.5 and 1 x 10/4 = 2.5. Thus, you can eliminate choices A and B. Because this is the least squares regression line, it must contain the point (xbar, ybar), which in this problem is (5, 10). So I just plugged in 5 for x in the remaining 3 choices, and only D will give you y = 10 when x = 5. So D is the answer.</p>
<p>so…when finding confidence intervals for proportions, it’s p^ +/- z x sqrt[(p^(1-p^)/n)]</p>
<p>but for the z test for proportions, it’s just z=p^-p/sqrt[(p(1-p)/n)]</p>
<p>…as in, no 'p^'s in the latter? which makes no sense because for means, the 2 things are the same…</p>
<p>^ I think you use p-hat when you’re doing a significance test for the difference of two proportions?</p>
<p>guesswhoooo, for the z test you use the population proportion because you are trying to do a z test for the null hypothesis that p hat = p population.</p>
<p>For the two sample proportion test, you do use p hats.</p>
<p>just got done…wasn’t that bad.</p>
<p>curse my last minute confusion. that really wasn’t that bad. it could have been so much worse. i def got my 3 :)</p>
<p>If I completely skipped one of the FRQs, and did decently on the multiple choice, how well would I need to do on the other FRQs to get a 3 or 4? D:</p>
<p>Man, that MC was easy…
FR…a bit tougher…</p>
<p>It wasn’t great, but I think I have a small chance at scraping by with a five, probably a high four. Lack of time to finish one FRQ will probably kill me, but at least it wasn’t #6. Onto Calculus!</p>