Let's study for AP STAT!

<p>Ok, so I have no idea how well I am going to do on this test but I thought maybe we could all just put random info to review.</p>

<p>I will start with hypothesis testing:
State assumptions: SRS, random sample, population is more than or equal to 10n</p>

<p>State Ho and Ha: equation than written statement, want to disprove Ho</p>

<p>Select test statistic: one sample z/t test left or right tail etc.</p>

<p>Calculate: I cant write this because there are two many symbols</p>

<p>Reject or Accept</p>

<p>State conclusion in terms of problem</p>

<p>oh, and incase anyone did not know, if you have a problem remembering how to put in certain info into you calc to get a certain equations, you can do the following:</p>

<p>you must have a ti 83 plus or higher:
go to apps and go to catelog help
click catelog help until it goes to a blank screen
go to the equation (for instance binomcdf) and rather than hitting enter, hit the "+" on your calc
this will bring up the symbols of what is necesary for that equation
you can leave the help on all the time without any problems</p>

<p>Umm I don't have "catelog help"</p>

<p>if you go onto the the TI site you might be able to get it. I know ti84 has it on there already as well as the ti83 silver but I guess not all.</p>

<p>The standard formula for the z-test, t-test and basically everything else excluding the chi-squared test is:
(average of statistic minus average of parameter) / (standard deviation of statistic)</p>

<p>If anyone could go over the two sample testing for T and Z testing, I would be overly grateful.</p>

<p>If another person could go over when to use the Z test, when to use the T test, when to use the Z or T two sample test, and Chi-squared Test, my gratefulness would double its amount :D</p>


<p>and for the above post, what exactly did you mean by: "catelog help"?</p>

<p>under my apps it is labeled "CtlgHelp"</p>

<p>sometimes you can use both the Z and the T test, but to really differentiate you must know the standard deviation of the population. if you know that, then you do a Z test, but if not, then you do a T test, which means that most of the time you should be doing a T test since knowing the standard deviation of the population is highly urnealistic.</p>

<p>I thought the main difference was that z tests have large samples while t tests have to be used for small samples</p>

<p>Rule of thumb is if standard deviation of population is unknown or sample size < 30 then do a T test</p>

<p>Yay for stat "studying". Anyone know how much room there is for work on the actual exam? I'm going through Barrons and can barely fit anything...if this is realistic my graders are going to have one hell of a time.
Quick question: what's the difference between stratification and blocking? And how would you figure out the margin of error if you don't know the standard deviation?</p>

<p>stratification i believe, is used when it is an observation.
blocking is for experiments.
if you are not given the standard deviation, chances are that you will be given the N, the number of trials, and P, the probability, thus you will be able to use the formula, squareroot of ( P times (1-P) divided by N), in order to get the standard deviation.</p>

<p>For the above equation that I have mentioned, does anyone know if the standard deviation given by the formula refers to the population or the SRS?</p>

<p>Keep em coming!</p>

<p>if you arent given a p, then you assume that it is 1/2 because that is the maximum for p(1-p). One question asked how it changes if you then find out the p.
It says that the standard deviation of the set of sample proportions is roughly that. So if you were to make a z-interval on the 83, would sigma just be sqrt(p(1-p)) and n = whatever it is. And then how do you get a binomial z interval for the mean.</p>

<p>hm.. binomial z interval for the mean.. I can't think of any cases where I've seen that come up before.. and therefore I am clueless about that subject :( sorry!
But I'd think it would be approximately the same for binomial as it would be for the normal ones. In any case, if you are unsure, use the calculator function T-test intervals under Test under Stat, or simply leave the answer choice blank :D</p>

<p>does anyone have any practice stat exams with answers provided with them?
I have the questions but I am wondering how the graders want the answers formatted.

<p>Does anyone know how the level of difficulty of barrons compares with that of the actual test? Also, where can I get previous years' exams?</p>

<p>I'm guessing that Barron MC is a harder.
And the FRQ is slightly harder.</p>

<p>t test is when you do not know sigma (most of the time).</p>

<p>Is the PR book worth me going to barnes and noble tonight and cramming?</p>

<p>No, but the Barrons might be, depending upon your motivation level and the score you want. And on how late they're open.</p>

<p>To keep this on topic...what's less expensive, a SRS or a stratified poll?</p>

<p>I would say the two wouldn't make much difference, since the only difference would be observations of the whole population at once, opposed to observation of the whole population in seperate stratas. But if we do not ignore the slightly more costly attempt to stratify the population, I would say the stratified samplings would cost a Tiny bit more than the SRS, not that it would make much difference.</p>

<p>Can anyone tell me if there will be any two sample t/z tests on the free response? (was it part of the AP Stat curriculum)?</p>