For instance, the SATs. Unless they’ve changed since I took them, you got 400 out of 1600 points just for showing up. And there’s a commercial for one of those free credit report websites that implies credit scores run between 450 and 850.
Why? What’s the reasoning behind not starting a scoring model at 0?
Often these types of systems are deisgned to account for the people who don’t take the test. For example if 40% of people in an age bracket don’t even bother to sit the SATs they immediately account for scores 0-400. The lowest range is reserved for the non-attendees so the sample space for the results effectively becomes the entire population, not just those who were tested.
That can make the results easier to manipulate statistically, especially when you have to deal with equivalencies. For example if someone applies for a college 15 years after they graduate the college can attempt to correlate his SAT score with the current intake more easily if they already know that it is tsandradised to population size because they dont; need ot account for numerous social trends. Or a foreign student with no SAT can be more easily compared based on other factors if they already know what percentage of people are in the US don’t have an SAT.
I think you can theoretically get 0 by answering every single question wrong since you lose 1/4 marks for a wrong answer. Other scales, although theoretically from 0 - 100, seemed to have morphed into a tight band. Wine scoring, for example, rarely goes below 80. Debating, at least in Australia is scored from 75 to 85.
LET NITPICK=ON
Just for the record, the main scoring system for credit reports has:
0
OR
450-850
If you’re a college freshman applying for a credit card at the little American Express stand in the student union, you may actually have a 0 score. It’s reserved for people with very, very thin files.
OK, so why start at 450? Why not 100 or 375 or a million and six? It seems so completely arbitrary.
Actually, they just changed it. The analogies are gone, and there is now a writing section that receives its own 200-800 score, so the base is now 600 and the max, 2400.
Good question.
Especially good considering the statistical validity of the model. The difference in default risk between a 615 and a 625 is just not that big.
It would probably work just fine to divide by ten:
0 or
45 to 85, with 1-point increments.
I’m not sure of the value of reporting that I have a 682 verus that I have a 680 when the system probably has at margin of error around 10-30 points depending on who you ask… it definitely doesn’t have precision to the point where a 682 and a 683 have any meaningful differences.
The Law School Admission Test (LSAT) is similar. You get a score of 120 just for putting your name on the answer sheet. The top score is 180.
It seems to me it would be easier to rank the thousands of people who take each test if there were 180 points instead of 60.
A lot of things about any rating are arbitrary, including starting at zero.
A few points.
Admission test scores will be interpreted by schools who know what the numbers mean. People with low scores can often rewrite the test. These tests are often lucrative hoops – it costs a fair chunk of change to write them. If you do badly, a score of 120 is arguably less demeaning than a score of 0. It is arguably more demeaning to tell someone there credit score is 0.
Some scores are based on tradition. The Glasgow Coma Scale measures level of consciousness. Dead people score 3, most people are 15. This was the way the original reseaqrch was reported, so this is the way things are still done.
Arguably makes statistical analysis easier, but really it is no big deal to account for this. But there is no compelling reason to make things simpler, particularly where a lucrative monopoly exists.
Not if the person understands that 120 is the lowest possible score, which is helpfully explained in the information that comes with your test score.
An admission office’s ability to better differentiate and rank the thousands of law school applicants would seem to me to be a compelling reason. As it is, raw LSAT scores of 1 to 13 translate to a reported score of 120, while raw LSAT scores of 98-100 translate to a reported score of 180. Between those extremes, some reported scores equal one raw score (e.g., 154 = 67), while other reported scores equal a range of two raw scores (e.g., 166 = 85 or 86).
I’ll attempt to make a case the Golf fits this model.
In one way; You can’t get a score of zero on any individual hole or ‘round’ (typically 9 or 18 holes)
In another way; Each golf hole is rated at a ‘par’ of (typically) 3 or 4 or 5 and you can go even (‘par’), above, or below. ‘Par’ itself is not considered a ‘zero’. On Televised matches I believe (not certain) they show cumulative scores as + and - and ‘even’ but no ‘0’.
You may write the LSAT and understand your score of 120 is bad, but most folks wouldn’t know that. Is it their business?
If admissions offices were genuinely unhappy with their ability to interpret test scores and rank applicants, these tests would be done differently. If someone scores 166, presumably other criteria would be used to determine the better candidate and this would matter more than a test score of 85,or 86. Different schools weight the MCAT differently – McMaster and Ottawa are good Canadian medical schools that don’t require it at all!
I think you should read “re-take” in place of rewrite.
I don’t see a big difference between “retake” and “rewrite”, but you may be right. Rewrite usually means “a revision”, but can also mean “to write again”. Retake is less ambiguous.
I don’t think that the expression to “write an exam” or “rewrite an exam” is as commonly used in the U.S. as it is where you are, Dr. Paprika.
I’ve never heard of a student “writing” an exam in the U.S.A. The expression is always “taking” an exam.
To me, those mean exactly the same thing.
“Writing” an exam is a bit more common in the UK and South Africa, I think.
Here we tend to “do” an exam, or “sit for” an exam.
In the UK at least, exams are written by the examiners (or the educational body or whoever). They are taken or sat by the student.
Most online survey type things don’t allow a zero score - you know, “rate this article” or “rate this picture” are normally 1-5 or 1-10. This isn’t just online surveys, in fact. This means that whatever it is that is being ranked can’t have a score of 0 (unless no-one’s voted for it yet, I suppose).
That’s all I got.