<p>Is there a curve on the ACT? If so, what is a general curve that is usually given?</p>
<p>My scores on practice tests from the real act books are as follows:
30 English, 34 Math, 29 cr, and 28 science reasoning...would my actual score be very similar to this?</p>
<p>Is it true that people tend to do a lot better on the ACT vs. SAT?</p>
<p>And, lastly, I've been looking at a lot of college Act composite ranges and i noted that even some ivy league schools such as cornell have a score composite range from 27-30....why is this so low, condering that its an ivy?</p>
<p>Yes, there is a curve. The ACT is scored so as to have a normal distribution, as is the SAT. Here is a chart showing which percentiles certain scores are. <a href="http://www.actstudent.org/scores/norms1.html%5B/url%5D">http://www.actstudent.org/scores/norms1.html</a>
So, a range of 27 to 30 would be the 90th to 97th percentiles of testtakers. I wouldn't be surprised if the SAT composite range reflects a similar percentile range.</p>
<p>Your practice test scores would probably be close to your actual score, give or take a little. On any given administration of the test, one's actual score will fall within a range, all else remaining equal.</p>
<p>No, it isn't generallly true that people do a lot better on the ACT. SOME do and some don't. Some do better on the SAT and some do the same on both. If you google on "ACT SAT conversion chart" you should find how scores compare. This is based on an actual study of people who took both tests and is how colleges convert scores so applicants can be compared.</p>
<p>Since both tests are curved to a normal distribution and are offered to similar populations, people can't generally do better on one or the other. In other words, only 1 percent of testtakers for either test will score at the 99th percentile, etc.</p>
<p>The ACT is not "graded on a curve". Go here: <a href="http://www.actstudent.org/testprep/index.html%5B/url%5D">http://www.actstudent.org/testprep/index.html</a> and look at the pdf "Preparing for the ACT". On page 63 is a table for converting raw scores to standard scores for the practice test in that booklet. Each test form has a table that looks like this. The values are different depending on how hard the test is. The conversion table is determined once for each test form by administering it and a test form of "known" difficulty to equivalent populations at the same time. If they do better on the new test, then the new test is likely easier, so they make the conversion harder so that the two tests are equivalent. It has nothing to do with percentiles. This all happens long before that test form is administered to you. So it's possible that, for example, no one would get a 36 in English for a particular test date if no one got all the English questions right. They do NOT curve it so that those who only missed one question get a 36, unless the conversion table was set up that way in the first place.</p>
<p>The percentile rank table Diane directs you to is created each year. It is based on testers from LAST school year (04-05) or maybe last year and the two previous years. So if you score in the 75th percentile, it means that 75 percent of last year's testers got your score or lower. It has nothing to do with the population you test with.</p>
<p>strok3s, the composite range you mention is probably for the middle 50 percent. That means that if you ignore the bottom quarter and the top quarter of Cornell students, the remaining students scored 27-30.</p>
<p>Thanks for the clarification. I had not seen this explained before. But isn't the "test of 'known' difficulty" -- to which a new test is made equivalent -- one that has its raw scores converted to standard scores that are normally distributed, i.e., curved? Is it really likely that one administration of the test would yield scores that aren't distributed in this fashion -- that, say, someone scoring a 30 would NOT be roughly at the 97th percentile of those taking that particular test? You've added a technical detail, but I don't know that in practice whether it makes much difference to a testtaker. It is theoretically possible that the distribution of scores would be different, but it hardly seems at all likely. </p>
<p>My understanding came from a discussion of the SAT and the ACT in an introductory statistics book in its section on normal distributions. I can't see why you say it has nothing to do with percentiles. The chart shows percentiles and the scores show a normal distribution (what I call a "curve"). Why is the 97th percentile at 30? Because you are dealing with an original mean of 18 and a standard deviation of 6. Two standard deviations above the mean is 30 -- and two standard deviations above the mean in a normal distribution is the 97th percentile. </p>
<p>Are you saying that ACT scores ever do NOT fall in a normal distribution?</p>
<p>It is true that a lot of ACT composite scores do not match up to what their converted SAT scores are in college data. This is because there is usually a much smaller sample taking the ACT test at these colleges. Also, some colleges let you submit ACT instead of SAT or Sat2, so a lot of athletes take the ACT; in addition, a lot more athletes take ACT as well, and their impact gets more pronounced when you have a smaller sample. It's still a primarily midwestern test, and it's just a fact that people on the coasts take standardized testing and academics a lot more seriously.</p>
<p>If you're a non-hooked, non-legacy, non-donor, non-minority, you'll definitely need at least a 29-30 to have a shot at cornell and other ivies.</p>
<p>Convert your ACT scores to SAT scores using collegeboard charts to see how good your score is. That is what colleges do.</p>
<p>Diane, the score conversion for the "known" test was created in the same way, and does not necessarily have (and probably does not have) a normal distribution. The percentile ranks you've refered to are assigned after the converted score is determined; your earlier explanations assume the converted scores are assigned based on the percentile ranks. It is not true that 1% of scores are at each percentile; and, in fact, if this year's testers are really terrible compared to last year's, it's theoretically possible everyone who tests this year would get a rank in the bottom 50% (i.e., no one would score above 20). If testers get better and better each year, average ACT scores will rise.</p>
<p>There must be an "original" score conversion that was set when ACT first started doing score conversion and test form equating this way (which I think was in 1989), and it seems possible that a normal distribution was used to determine that. The only reason the conversion would look anything like a normal distribution now is if the abilities of testers hadn't changed over the years. But it's clear that the current percentile ranks do NOT reflect a normal distribution--it starts to get WAY off as you drop below a 28 or so.</p>
<p>The raw score is converted into a scale score using a table, and the percentile rank is then determined from the scale score using a different table. Both tables are created long before you test. So the percentile distribution does not affect the score you will receive. And, moreover, your score is unrelated to the scores of the others who tested at the same time or in the same year. If you bombed this April, it doesn't matter if everyone else bombed too. </p>
<p>The whole point of the way different test forms are equated is that a score of 30 today means the same thing as a score of 30 ten years ago--it reflects the same level of educational development--even if a 30 is at the 97th perecentile today versus the 90th percentile ten years ago. Its intent is actually the opposite of the sort of "grading on a curve" that makes your score meaningful only relative to your peers--which is what I believe testers are talking about when they ask "Is the ACT curved?"</p>
<p>Mrs. Ferguson, thanks for the explanation. It obviously is not a simple thing to explain. </p>
<p>Your definition of "curved" as dependent on how people do on that very test may be what people mean when they ask. If they mean, "Can a majority of folks score 30 and above?" then I think the answer is theoretically yes, but realistically no.</p>
<p>I'm surprised that the scores aren't normally distributed anymore. I can see how if you set things to equal previous tests, the "copy of a copy of a copy" phenomenon could get things a little out of whack. But I would think that students' abilities would still be normally distributed and thus a test would reflect that.</p>
<p>Do you happen to have a link that shows the number of people with each score recently and in past years? Don't go to any extra effort; I would just find that interesting if you know of one.</p>
<p>Thanks again for taking the time to explain all this. My understanding was obviously a theoretical one based on my statistics book and the fact that my daughter's percentile and score happened to fit the normal distribution pattern explained in the book. It didn't go into the question of making the test consistent from year to year, which I can see would complicate things.</p>