The MORS system is an ABC grading system where ABC are the grades for performaces falling on the main sequence of the bell curve, F is for failure (low end outliers of the curve) and O is for Outstanding (high end outliers of the curve). ABC grades can also get pluses and minuses. It's that easy.
Using movies as an example, on this system an O is a five star movie. An F is a zero star movie. An A+ is 4 1/2 stars, and a C- is 1/2 star. So, an A+ movie is an excellent movie, one clearly worth seeing -- it's one with excellent performances, directing, cinematography, and script. An O exceeds this in that it is also a classic, a very rare entry into the pantheon of the canonical films. Any movie less than a C- was a disaster: a mistake to see under any circumstances and one that We believe should never have been made or released. That's an F, a failure. In practice MORS is common sensical and easy to apply.
But why, for Heaven's sake, why? Well, let's discuss (by which I mean I am going to just tell you).
We note that there are five rating system's commonly used in the United States (from where I write).
The first is the ABCDF system, which uses pluses and minuses except for F which means failure and is always just an F. Sometimes people say it's bell curved and sometimes not. When it is, like in a schoolroom, putting it on a bell curve is often unfair because (a) the sample size -- a class of usually 30 or less -- is too small for a bell curve; (b) no grading standard between classes and schools exists, making application arbitrary; and (c) since not all classes are on a bell curve and students on a bell curve may appear to have done worse than students on a straight scale when they performed just as well. (That last one I've put in very vernacular terms, but just extrapolate.) When grading isn't on a bell curve, however, it fails to recognize that there may be a substantial difference between, for instance, students earning A's since some may be way out from the norm -- more than a standard deviation or even two -- and some only a little. In reality a substantially large population performs on something close to a bell curve (sometimes skewed up or down, depending on the material). It would be preferable if the system could reflect reality while avoiding (at least to an extent) being arbitrary. On the ABCDF system described above there are 13 possible grades (assuming A+ is possible), and F is sort of a weird grade designation because it is connotative and can't be modified but there's no parallel on the opposite end of the spectrum. Yet letter grade systems are appealing because they are so common and have an easy point of reference for all of us.
The second is the star rating system, using up to four or five stars depending on the system and usually using half stars. One can see how that fairly parallels the ABCDF system. It's usually linear, and zero stars may or may not be an option. It's nice and visual, which is appealing. Assuming a five star max system with half stars and zero stars possible, there are 11 possible ratings.
The third rating system is a tiered numerical system, often 0 to 4 or 4.33. This corresponds to the ABCDF system, with 0 equaling an F, 4 an A, and 4.33 an A+. Each +, -, or whole grade increments at .33 with the letter grades as whole numbers. This system is averaged as linear regardless of whether graded by each grader in a linear fashion. It has 13 possible grades. A not too uncommon variant of this system is to abandon the fractions and grade from 0-10.
The fourth rating system is percentages (which are not to be confused with percentile). This typically reflects a percentage of questions correct or the like. It is often correlated to a letter grade (so A typically means 90% of the questions were answered correctly, B 80% of the questions were answered correctly, and so forth). but the correlation is strained because it is not clear how the pluses and minuses fit into this system and where an F goes. So, for instance, if a 90% is an A and an 83 is a B+, then what is, say, a grade above, 96? Also, presumably anything below 60% is an F (though sometimes it's below 50% or 55%, depending on how D- are treated) and is usually automatically reduced to 0. It's not clear why a student who gets 47% right should be treated the same as one who got them all wrong.
The fifth and final rating system is a percentile system which simply measures performance compared to a large pool and thus takes into account the bell curve effects of performance. This is a great system if the grading system is objective and the pool is large enough. But, alas, neither is usually the case.
None of these systems take into account -- at least on their surface -- improvement between grades (and whether that should affect final grades), learning effort versus knowledge versus understanding and insight and how to grade for each, and test effort and performance issues that may substantially affect outcome. Those are serious educational concerns, but not my concerns.
The MORS system takes advantage of the fairly linear distribution of performances clustered in the center of the bell curve while recognizing there are outliers at the top as well as the bottom.
"But wait, Galileo," you say, "just wait a sec. You haven't told us a numerical measure under the MORS." That's true, because numerical systems suck, as you can see above, for small pools. But if we must convert MORS to numbers, then we use a 0-5 system where O=5 and F=0. There's perhaps some marginal points we might award to F efforts that can't per se be used as torture (those points being 1.33 or less), and we would recognize an O- at 4.67 for the outstanding efforts that are just not as canonical as if, say it was a film, Lawrence of Arabia. We do this so that A's can still earn a 4, B's 3, and C's 2 to correlate with other numerical scales. a 5 is rare but exists in theory. It's like having an amp that goes to 11.
See? It's just that easy.
No comments:
Post a Comment