<p>We didn’t really review in my class, we just did all the released multiple choices and FRQs beginning in early april.</p>
<p>I took a class. Though, due to senioritis, we pretty much just messed around for two or so months.</p>
<p>
[QUOTE=miracleboy]
We need to know transformations for the AP test? i haven’t really seen this on a MC before.
[/quote]
</p>
<p>Pretty sure you do. Given a scatterplot, you need to be able to tell if the correlation is linear or non-linear. If it is linear, well good for you…do the standard old linear regression stuff.</p>
<p>If its not linear, you should be able to tell if its a power model, exponential model or a logarithmic model. You can get the equation from the calculator and you convert the exponential model (x vs y) to (x vs log y); this is done to make the model linear.
To do the Power one, you need to convert (x vs y) to (log x vs log y), same reason as previous.
The whole log transformations is done with logging both sides / raising both sides to 10…all that fun stuff.</p>
<p>Chances of it being on the MC section is unlikely, however, it can show up on the free-response, and the way I mentioned above is how you would do it.
The later parts of the question might ask you to sketch residual plots and comment on it I guess…not too bad with those.</p>
<p>I’m taking the class online, and I haven’t quite finished yet. Chi-square tests seem to be the only things left.</p>
<p>don’t forget linear regression hypothesis tests. i studied that last i recall</p>
<p>Study the chi-square tests! They’re important.
Also, a tip to everyone:
Make sure you state conditions/assumptions before running any test or confidence interval or else you won’t get full credit.</p>
<p>Can anyone explain a situation where we would need to pool the data?</p>
<p>I just checked, and the last two topics are expected values and independence and inference for least squares lines. I guess I’ll just watch the videos and do a few practice problems.</p>
<p>I also need to read over tips for the FR.</p>
<p>Pool the data when the population standard deviations are equal or when using proportions.</p>
<p>
[QUOTE=NYEM]
Can anyone explain a situation where we would need to pool the data?
[/quote]
</p>
<p>You pool it when the variances are equal. If they do not say anything about the variances, you ALWAYS ASSUME they are equal. If they give you the corresponding standard deviations / variances or tell you that ‘x is more spread out than y’ or something like that, then you don’t pool the data.</p>
<p>Someone remind me the format for invNorm( on the calculator?
You need to calculate the z-score by hand before using invNorm, right?</p>
<p>invNorm(area under curve, mean, std dev)</p>
<p>i think invNorm is supposed to give you the z score</p>
<p>So what would you enter for area under the curve?
I thought you enter invNorm(critical t or z, mean, st.dev)…</p>
<p>invNorm(area, mean, standard deviation) I think. I’m using a ti-89 so it might be different for you.</p>
<p>“On average” lol, how many problems can you get wrong in the multiple choice section and still get a 5?</p>
<p>idk about how many questions you need right to get a 5, but according to barrons
if you get 20 for your raw score in the mc and 2/4 on all the free response questions, you get a 4
my teacher told me that they grade the free responeses really easily and that if you have anything relevant you will get at least 2
but im going to bomb this anyway
i havent learned a thing all year</p>
<p>anyone have a good sheet that summarizes all of the inference tests + assumptions? It’d be helpful to study off of.</p>
<p>Also, will calculator memory be deleted before the test?</p>
<p>second videogames question… and also, raw score of 20 means 20 wrong on MC out of 40? or 15 wrong out of 40? I really need a 5.</p>