<p>I want to reveal some very startling info. I added up the difficulty of the 2007-2008 sample test in CR and got a number of 203. I did the same thing for the 2008-2009 test and the result was 200. Now it is not a huge difference, but the 2008-2009 test still had a slightly better curve.</p>
<p>WHY???</p>
<p>Because the questions were harder. If there was a difference in difficulty of over ~5-10, the test would no longer be standardized.</p>
<p>no the CR questions in the 2007-2008 were harder...meaning it should have had a better curve.</p>
<p>There ARE test dates with better curves, but trying to predict them is just stupid. Pick whatever test date is convenient for you and just take the test.</p>
<p>Unfortunately for you, that's not conclusive proof at all. Statistics teaches us that if you sample size is only one or two, then your results might as well be outliers. Even if we ignore that, rockermc is right in that the difference in difficulty is not significant at all. Finally, these were sample tests you did calculations on, right? Sample tests mean nothing; do calculations on real, past exams and then report your findings.</p>
<p>the sample tests ARE real past ones vrdabomb. I can see where you are coming from mamamia, but honestly 3 points is hardly any difference. besides, the 1-5 ratings are subjective, I've seen 3 point problems harder than 5 pointers and what not. So the moral of the story is...It doesn't matter. Take it when you are most prepared.</p>
<p>as i've repeated again and again on these forums, test dates with better curves apply only to those people at the very top (i.e. close to 2400). so if you're at 200-203, i'd advise you to stop worrying about curves and start getting better first.</p>
<p>jamesford is right on the money, but that having been said, another "explanation" here is that the difficulty numbers of 1-5 do not reflect very easy or very hard difficulty levels. I.e., a couple of extremely hard question may be enough to make the curve a little nicer, but these are still just 5's, compared to another test with two "typical" hard questions.</p>
<p>well if you consider ~750 CR to be at the top, then I am srrinath. but w/e, this discussion is useless. As yettiddqq8 and fignewton pointed out, curve data is more complex than I thought.</p>
<p>yeah... really conclusive.</p>
<p>no i don't consider 750 to be high enough that the curve affects you. generally 750 is a -5. it only matters when the number of questions you generally get wrong is 800 on one curve and non-800 on the others.</p>
<p>The concept of "the curve" is designed to make very test equal in difficulty. The harder the test, the more lenient the correlation between raw score and scaled score.</p>
<p>However, the flaw of this, is that people who are scoring above average don't find the same type of questions difficult as those scoring below average. And those scoring in the 2100, 2200, 2300 range mostly make errors due to carelessness, which does not have anything to do with the "difficulty" of the test.</p>
<p>I've compared the curves from about 10 past SATs, and found that they're almost identical once you get below the 700 mark. The only time the curve really comes into play is when you plan to get -1.25 (1) in Math or -3.75 (3) in CR. </p>
<p>-1.25 Math can be anywhere between 770 and 800
-3.75 CR can be 760 to 790</p>
<p>As to the business of adding up the little 1-5 ratings of difficulty. . . .</p>
<p>I think that if, say, you get 54/54 on math when the curve is -1 = 780, and someone else gets an 800 when -1 = 760, the latter score is better and thus unfair to the first test taker.</p>
<p>^Why unfair? Both will get 800s; the only way to separate these two is with more questions (5 hour SAT, anyone?)</p>
<p>If you want to try to find proof of better test dates, a survey of released SAT math curves can be found here</a> (pdf file) ...</p>
<p>well in that case there are better test dates, but you don't know which dates there are. There are no clear cut months with better tests so there is no point thinking about this. Whatever test you get is the test you get.</p>
<p>Well, both will get 800s, but the one with -1 = 780 had more hard questions to answer in order to get the 800.</p>
<p>True, but you have to stop somewhere ... consider two people in the same math class, one very smart, the 2nd brilliant. Each takes a test and gets 100%. Is this unfair to the brilliant person? Possibly, but the only way to give the brilliant person "justice" is with a longer test, more tests, etc.</p>
<p>well in that respect it is only unfair to the ones getting an 800. If it was anything else, the curve would handle it. but if you get an 800 you should be happy regardless of whether its unfair. in fact i would actually be proud to have persevered a tougher test to get the 800. its like playing halo. do you feel more proud after winning legendary or after winning easy?</p>
<p>I guess that's why they have those AMC contests, I suppose. Thooose have a high enough ceiling (even if it didn't, the AIME would take care of that :P)</p>