No more deductions for wrong answers - Four Questions

Test Preparation AP Test Preparation
21

<p>Hmm, the 2009 Euro history curve is quite interesting. If we look at the cutoff for a 5, it is 114 with the guessing penalty, but then 119 without the guessing penalty. If we assume that people get the same amount of points taken off for the MC as they do for the essay, then we have (180-119)/2 = 30.5 points on average taken off for the MC for the curve without the guessing penalty.</p>

<p>If we then factor IN the guessing penalty, we get 30.5 + 0.25 * 30.5 = 38.125 points taken off. Add this to the 30.5 points taken off from the FRQ portion, and we have a grand total of 68.625 points taken off over the whole test, or a 111.375 cutoff to get a 5. If we compare this to the test with the guessing penalty to begin with, we see that this is actually LOWER than the 114 cutoff that was originally with the guessing penalty.</p>

<p>So what does this mean?</p>

<p>If we look at only the raw data, this means that the test is actually more lenient WITHOUT the guessing penalty. </p>

<p>What about my theory?</p>

<p>Well, two things. Either 1: The guessing theory that I posted earlier only accounts for some of the changes, but there are more important factors that outweigh that theory. Or 2: My original assumption that people, on average, get the same number of points taken off from the essay as they do from the MC is incorrect.</p>

<p>Anyone have any discussion on this? I only have one brain, so I could easily be missing something that you guys may have already known - if so, share it to everyone and I will readjust accordingly!</p>

22

<p>I agree with you, SeekingUni, in that the percentages of 4’s and 5’s will most likely stay the same. However, I am confused at how my math shows the curve to be more lenient, but my logic to show it to be harsher. Of course, this is probably trivial to most of you and slightly off-topic from the OP, but I am still curious!</p>

23

<p>I think that the adjusted curves need to be taken with a grain of salt (if that’s the proper idiom). Unless CB took the MC data from 2009 (which I doubt they did or they even have) and ran everything through with essay and adjusted MC score and then had a person determine the curve, the adjusted curve is nothing but an estimate. Sure, it’s the best we’ve got for predicting what percentage is what score, but I don’t know if it’s just a CB guess or a curve re-created from re-tabulated data.</p>

24

<p>It’s an obvious fact that people will generally do better without the guessing penalty.</p>

<p>With the penalty, you have a 1/5 chance of getting a question correct (since there are 5 choices). So that means, in 5 questions, you would theoretically get one of them right. If you got one right and the other four wrong, that’s [1 - (1/4)*4] which is 0. That’s a fair score, because if you guessed on them, you didn’t really know the answer to any of them and don’t deserve any points.</p>

<p>Without the penalty, you would keep that 1 point for every 5 questions. This means you could guess on an entire 100-question exam and would theoretically get 20 points, whereas with the penalty, you’d get 0. It doesn’t seem like a huge difference, but it can add up. </p>

<p>Without the penalty, you are likely to spend less time on each question as well… because then you don’t have to worry about making sure you pick the right answer or if you should even pick an answer at all. </p>

<p>So as a result of all of this, scores are bound to increase.</p>

<p>The curve is set each year to distribute the scores a certain way so that there is a certain percentage of each score (1, 2, 3, 4, 5). As the test scores change, so will the curve. Everybody says that the curve will be “harder”, but in reality, it’s no different because the test itself is a bit easier, giving you the opportunity to earn more points than you would with the penalty.</p>

<p>It all balances out. You probably won’t see a significant difference in anything, to be honest.</p>

<p>You also don’t know if the difficulty of the questions is changing. Perhaps they are increasing the difficulty, making for an opposing force to this lack of a penalty.</p>

25

<p>

</p>

<p>That would make a lot of sense, now that I think about it. They’d have to do a lot of math and calculations if they adjusted EVERY curve precisely to fit the new system. </p>

<p>

</p>

<p>That is in fact something I did not even think about! Yes, I think it all makes sense now. I did not consider any psychological aspects, nor did I consider College Board’s point of view. Thank you both for clearing up my confusion!</p>

26

<p>Well, they only have to re-calculate curves for released exams, which makes it a bit simpler, I suppose. Also, if they have all the data in a computer, maybe they could run it through a different formula and have a person determine 5/4/3/2/1 cutoffs as always.</p>

27

<p>Very interesting analysis. So the odds are that the percentage of 5s and 4s will remain the same. So what prompted the CB to make this change? So far the only theory has been that this is a marketing strategy designed to compete with the ACT. But wait. The ACT does not offer AP tests. Maybe this is a preliminary move to eliminating the “guessing penalty” on the SAT. What do you guys think?</p>

28

<p>Hmm, if that was true, why start with the AP tests then? Isn’t the AP test given out on a much larger scale than the SAT? Why not just go directly to the SAT?</p>

29

<p>

</p>

<p>The SAT is taken by more people.</p>

30

<p>Yes, but I mean frequency. Are there more AP tests given out than SAT tests? (For example, many people will end up taking 5+ AP tests while taking only take 1 SAT test)</p>

31

<p>

</p>

<p>This is just not true. Similar to the SAT percentiles, there are no AP score percentile “targets”. The fraction of test takers scoring an 800 on the SATs varies solely according to the quality of the test takers. The CB does not decide that, say, 0.5% of SAT students will get an 800, or even that 50% will score more than 500. In the same way, the percentage of AP students getting a “5” or any other score is not fixed to a particular number. For certain exams, the AP score percentiles have changed significantly in the last decade (you can see this on apcentral by looking at score distributions for a given test over the last ten years).</p>

<p>It is likely that the percentiles will stay the same, but only because from one year to the next, the AP student population doesn’t change that much.</p>

32

<p>

</p>

<p>That’s not entirely accurate. From CB:</p>

<p>

</p>

<p>Src: [AP</a> Exam Scores](<a href=“http://professionals.collegeboard.com/higher-ed/placement/ap/exam/grades]AP”>http://professionals.collegeboard.com/higher-ed/placement/ap/exam/grades), [AP</a> Central - Statistical Information Based on Common Items](<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)</p>

<p>Wikipedia supplement:</p>

<p>

</p>

<p>Although they don’t go “we will have exactly 10.86% 5’s this year,” there is definitely a decision made on how to do the cutoffs. Bio from 2007 to 2008 nearly doubled in the # of 1’s because they adjusted the cutoffs for the exam, not because the people got dumber.</p>

33

<p>I wasn’t suggesting that cutoff scores are not adjusted. They are, of course, but not for reasons having anything to do with percentiles. Nothing in the CB material quoted above suggests that there are percentile targets and, most importantly, the standardization of the test could not be maintained if there <em>were</em> such targets.</p>

<p>The cutoff scores are generally adjusted to account for variations in test difficulty (as determined by the equating questions + readers etc.). In the unusual, somewhat controversial, but arguably necessary rescaling of the AP biology test in 2007/2008, the CB determined that the scores were not appropriately matched to college level abilities (the primary design goal of the AP scoring is that, for example, a “4” should be equivalent to a “B” in the appropriate college course). See [thread=542421]this thread[/thread] and the last post in particular for interesting details about the bio situation. The CB would not have reasoned that the percentage of students getting 1s needs to be 33% or so; they determined that a “2” on the 2007 was more like an “F” than a “D”.</p>

34

<p>

It was mentioned either in this thread or the other thread that one reason for the change may be because of the increasing rate of lower scores (1s and 2s).</p>

35

<p>where did you hear that they would stop the point deductions?</p>

36

<p>Look in earlier posts or in this thread: <a href=“http://talk.collegeconfidential.com/ap-tests-preparation/940263-2011-exams-no-more-deduction-wrong-answers.html[/url]”>http://talk.collegeconfidential.com/ap-tests-preparation/940263-2011-exams-no-more-deduction-wrong-answers.html&lt;/a&gt;&lt;/p&gt;