Race to the bottom: Red Grader, Blue Grader

[Aimless]

So is it fair to give Lewk a D- in statistics to hint at how far off the mark he is?


http://faculty.vassar.edu/lowry/ch1side2.html

[Khendraja'aro]

How big of a problem with Calculus? A grade will tell the individual how far off the mark he is. A 50 is a huge difference from a 15.

Say, do you actually read what I write? Here, I'll give you a hint:

*) Yes, it's a shortened version!

I'll say it again, your dumb. If there are three 10 question tests. I answer 7 correctly on one. That is a 70. Another one I answered 9 correctly. That is a 90. And yet another one I answered 8 correctly. That is a 80. I average those three scores together and I get an 80. I've made a B. This is done in pretty much every grade from 1 to 12 and most colleges as well. Your nonsense about blah blah blah not statistically meaningful is just that, blah blah blah.

"Your dumb"? That misspelling, my dear, suggests it's actually the other way around.

Regarding your example:
You can average the test scores. They're equidistant. They have a zero point.

However, in school you usually don't average the test scores. Your kind averages the grades you get for the individual tests. Not to mention that those tests are usually NOT multiple choice and thus subject to the problems I've outlined above.

Do you understand any of what I'm saying here? Because at the moment it doesn't seem like it.

[Nessus]

He mis-spells because he is better than the rest of us, and has special license to keep on trollin' by the moderation.

[Aimless]

A reluctance to rank people probably stems from some sad little fear of hurting people's feelings. Oh think of their self-esteem!

People are evaluated at their jobs. They are evaluated when they apply for credit. They are evaluated when they get an insurance quote. They are evaluated when they try to rent an apartment. Coddling people in school makes them ill prepared to deal with reality.

So what you're saying is that you feel compelled to rank people in inappropriate ways because you want to hurt them? Dude... dude, that's just awful.

Do you also want to prepare coddled children for rape, exploitation, physical and mental illness?

[Loki]

Oh, and by the way: You are actually not allowed to average grades. Statistics plain out forbids it since they're actually ordinal numbers.

That's a rather dishonest thing to say. It might be technically true, but we average plenty of other ordinal numbers. Doing so might violate some statistical assumptions, but we're not really trying to engage in statistical inference. One could also reasonably claim that grades are sufficiently close to interval numbers to be used as such.

[Enoch the Red]

He mis-spells because he is better than the rest of us, and has special license to keep on trollin' by the moderation.

I'd like to take this opportunity to point out that the proper spelling of Republican is not Pubbie.

I consider it remarkable that you too have been deemed better than the rest of us, and have special license to keep on trollin' by the powers that be.

[Khendraja'aro]

That's a rather dishonest thing to say. It might be technically true, but we average plenty of other ordinal numbers. Doing so might violate some statistical assumptions, but we're not really trying to engage in statistical inference. One could also reasonably claim that grades are sufficiently close to interval numbers to be used as such.

We are not? Grade statistics are used to distribute fundings. They're used to determine who gets into university. They're used to compare states. They're used to compare classes.

It's not me who's dishonest. Dishonesty comes from using those grade numbers for stuff they're not intended to. Mathematics is a tool. If you use those tools in the wrong way, you'll arrive at wrong conclusions. And no, they're NOT close to interval numbers.

Over here, to get a passing grade (4) you usually have to achieve at least 50% correct answers in a test. Which means that the failing grades 5 & 6 get the range of 0% to 49% while 1, 2, 3 and 4 get 50% to 100%. Not to mention that 1 to 4 are not evenly spaced again - for a 1 you usually have to achieve 95% or even more.

I'm not sure if you see it, but for me that's pretty far from "evenly spaced".

[Nessus]

I'd like to take this opportunity to point out that the proper spelling of Republican is not Pubbie.

I consider it remarkable that you too have been deemed better than the rest of us, and have special license to keep on trollin' by the powers that be.

Liberte, egalite, fraternite? NSDAP v Nazi? All of that, a smokescreen around Lewk's self-admitted trolling, to "see liberals squirm". If you think the monikers of Pubbie and Mittens are offensive, by all means, report me! Have you not already?

[Aimless]

Before we get completely side-tracked:

grades in the US is an important socially acceptable way to distinguish between eg. applicants for a job. They're valuable because they're easy, supposedly good enough proxies for various skills and traits, and above all socially accepted. Of course, we accept a lot of dubious things. The myth of the meritocracy, the American dream, the value of money, etc. Grades are like that. The social philosophical ethical psychological concerns etc are irrelevant.



The question then arises: is this a problem? Lewk for example may even enjoy the problems. Delicious unfairness! He may prefer to stick his... er... hand... in the sand. Wow that really doesn't sound right

[LittleFuzzy]

Hmmm. I think that this pronounced need to rank people is rooted in lower order thinking, in the genital area. I think it's from our monkey days. Or it could be a product of his upbringing. Or a product of extreme and unsettling uncertainty. It could be a lot of things. Swedes are notorious for being reluctant to rank people in many ways and I'm pretty sure their brains are about the same as Lewk's.

I might readily concede an evolutionary psych root for such mental activity. I don't give much credence to your claim that Swedes are reluctant to rank people though, given any substantive *vs shallow* meaning of the behavior. I can easily see how a cultural ethos against such activity can arise, I just don't believe it's actually being adhered to beyond the superficial.

[Aimless]

I dunno man, this place is a socialist haven for lazy unionists after all, and just look at the ridiculously low wages doctors make here compared to janitors, and all those years with a three-tiered-grading system at uni etc. There's a strong tradition here of not letting anyone think he's better than average in any meaningful way. It waxes and wanes. Physical appearance is pretty important. Partying, sailing, very cool. Social skills are extremely important. Likability. References, job experience. Grades? Often very unimportant, except possibly things like the bachelor's thesis.

Of course swedes judge each other. They're very fuzzy and inconsistent about it. Sometimes they're worse than Lewk. For the most part they don't think a low ranking should disqualify you from having a decent life.

[LittleFuzzy]

I dunno man, this place is a socialist haven for lazy unionists after all, and just look at the ridiculously low wages doctors make here compared to janitors, and all those years with a three-tiered-grading system at uni etc. There's a strong tradition here of not letting anyone think he's better than average in any meaningful way. It waxes and wanes. Physical appearance is pretty important. Partying, sailing, very cool. Social skills are extremely important. Likability. References, job experience. Grades? Often very unimportant, except possibly things like the bachelor's thesis.

Of course swedes judge each other. They're very fuzzy and inconsistent about it. Sometimes they're worse than Lewk. For the most part they don't think a low ranking should disqualify you from having a decent life.

You're studying medicine, right? I recall you talking about an ER rotation. Would you care to define "triage"? You can object that triage doesn't rank people but rather their immediate medical condition but an analogous objection can be made about what, say, Lewk's grades are ranking.

[Khendraja'aro]

But triage is more pass/fail than ranking

[LittleFuzzy]

But triage is more pass/fail than ranking

Triage can go rather further than pass/fail, but pass/fail is also a low-order ranking.

[Wraith]

Bug triages I particpate in are must/like/maybe/no.

[Aimless]

Triage is an excellent analogy. Triaging systems are fairly rough systems that, in emergencies, help decide who should be prioritised in order to prevent things like death. How good are these systems? Which parts are good and why? How do different systems compare with one another? How well do they serve patients who aren't at risk of dying as soon as you turn your back? Haven't the foggiest. In these respects grading is like triaging. We like triaging because it helps us make okay decisions in life-and-death situations. Try it at any other time, with or without refinements, and Americans start ranting about death panels and disconnected doctors.

Should education be treated as constant serious emergency? For whose benefit do we triage students? For tired employers and the best most desperate students. Overworked doctors and dying patients.

[GGT]

Triaging = prioritizing during emergencies. The medical-to-educational aspect doesn't compare very well. Medical emergencies are more time sensitive than educational emergencies. We're talking minutes or hours compared to days/weeks/months/years.

That's not to say educators don't "triage" by tagging some students by category, though. At Risk, Borderline, Struggling, Challenged, Failing, etc. Treating education as a constant serious emergency would burn out every teacher, and not benefit students who often move fluidly between categories with no special intervention at all. Sometimes all kids need is time and maturity.

[Loki]

It's not me who's dishonest. Dishonesty comes from using those grade numbers for stuff they're not intended to. Mathematics is a tool. If you use those tools in the wrong way, you'll arrive at wrong conclusions. And no, they're NOT close to interval numbers.

Over here, to get a passing grade (4) you usually have to achieve at least 50% correct answers in a test. Which means that the failing grades 5 & 6 get the range of 0% to 49% while 1, 2, 3 and 4 get 50% to 100%. Not to mention that 1 to 4 are not evenly spaced again - for a 1 you usually have to achieve 95% or even more.

I'm not sure if you see it, but for me that's pretty far from "evenly spaced".

If we stick strictly to statistical assumptions, then we have no way of getting an average grade. Even if the grade consists entirely of the number of questions a person got right, we can't assume that each question is of identical value (which in this case would mean difficulty I suppose). So we either pluck a grade out of a hat, or we violate a statistical assumption that is violated quite regularly, across all academic fields (in published journals).

Before we get completely side-tracked:

grades in the US is an important socially acceptable way to distinguish between eg. applicants for a job. They're valuable because they're easy, supposedly good enough proxies for various skills and traits, and above all socially accepted. Of course, we accept a lot of dubious things. The myth of the meritocracy, the American dream, the value of money, etc. Grades are like that. The social philosophical ethical psychological concerns etc are irrelevant.

As I mentioned earlier, that's actually not true. Other than grad schools and perhaps some consulting companies, no one will ask you for your college GPA. Even if they do, they'll be equally pleased with a B as they would be with an A. The main signal employers look for is not your GPA, but rather whether you have a college degree, and if so, from where.

[Enoch the Red]

Liberte, egalite, fraternite? NSDAP v Nazi? All of that, a smokescreen around Lewk's self-admitted trolling, to "see liberals squirm". If you think the monikers of Pubbie and Mittens are offensive, by all means, report me! Have you not already?

Never! Nor would I ever feel the need.

Of course, I'm not making the claim that spelling, or misspelling, constitutes trolling. I'm simply unable to imagine what kind of people would feel that constitutes abuse, especially when for that same person considers actual name-calling as a matter of course.

[LittleFuzzy]

Triage is an excellent analogy. Triaging systems are fairly rough systems that, in emergencies, help decide who should be prioritised in order to prevent things like death. How good are these systems? Which parts are good and why? How do different systems compare with one another? How well do they serve patients who aren't at risk of dying as soon as you turn your back? Haven't the foggiest. In these respects grading is like triaging. We like triaging because it helps us make okay decisions in life-and-death situations. Try it at any other time, with or without refinements, and Americans start ranting about death panels and disconnected doctors.

Should education be treated as constant serious emergency? For whose benefit do we triage students? For tired employers and the best most desperate students. Overworked doctors and dying patients.

It makes for a perfectly acceptable analogy, but I was really just using it as an accessible example of ranking which I know is used in Sweden. Using it as an analogy. . . our brains are constantly triaging incoming stimuli. We filter and rank, we prioritize, tag and classify, all the time. Even egalitarian, "non-ranking" Swedes. We do it unconsciously and we do it consciously, it's a basic pattern in how humans *and other complex life* think.

[Nessus]

Never! Nor would I ever feel the need.

Of course, I'm not making the claim that spelling, or misspelling, constitutes trolling. I'm simply unable to imagine what kind of people would feel that constitutes abuse, especially when for that same person considers actual name-calling as a matter of course.

It's not a feeling, it's what Lewkowski has named as his motivation. And the derision inherent in the term 'Pubbie' is no less than the one in 'Republican'; until the 1990s the Israeli sign language term for 'German' was the swastika. Pubbie is a 'bad name' because it stands for Republican, not because it sounds like pubes, or whatever. At least that has been my take on it, is the term rooted in something more offensive?

[GGT]

[...] until the 1990s the Israeli sign language term for 'German' was the swastika[...]

A new factoid I've learned today! That implies that after the 90's, Israeli sign language for 'German' became something else. What would that be? Is that what it is today?

[Khendraja'aro]

If we stick strictly to statistical assumptions, then we have no way of getting an average grade. Even if the grade consists entirely of the number of questions a person got right, we can't assume that each question is of identical value (which in this case would mean difficulty I suppose). So we either pluck a grade out of a hat, or we violate a statistical assumption that is violated quite regularly, across all academic fields (in published journals).

Ah, yes, because everyone does it wrong, it's okay to do it wrong as well. Right.

So, if everyone believes in a Flat Earth, it's okay to believe in a Flat Earth as well, evidence to the contrary, or what?

And, yes, doing an average is essentially plucking a grade out of the hat. You're making the same mistake again: Assuming that because you can do something, you actually have to do it because you don't get your precious data otherwise.

I'll remind you of your views next time you attack a study where someone did some statistical assumptions you don't agree with.

[Loki]

Statistical assumptions are regularly violated. We don't have unit homogeneity in virtually any study. There's usually temporal or spatial interdependence. Depending on the topic, there's auto-correlation. Omitted variable bias is nearly always present. So either you accept that not all statistical assumptions will be met and do the best you can, or you simply stop using stats.

[Nessus]

Stop being sensible Loki, it makes the backs of my eyeballs itch!

But really, it's a good point, people can be very sloppy with stats. Hell, I'm often sloppy on stats, because the trends are more important than getting the error bars or whatever just so. *shrug*

[Loki]

Want me to scratch that for you?

[Nessus]

Pussycats are all the same

[Khendraja'aro]

Statistical assumptions are regularly violated. We don't have unit homogeneity in virtually any study. There's usually temporal or spatial interdependence. Depending on the topic, there's auto-correlation. Omitted variable bias is nearly always present. So either you accept that not all statistical assumptions will be met and do the best you can, or you simply stop using stats.

However, Loki, in the case of grades NO statistical assumptions are met. You do not even TRY. There are no error bars. No standard deviations. No calculations of reliability. No nothing!

You simply calculate the average and make it out to be something that's reliable, set in stone and promptly people's livelihoods depend on an increment of 0.01 because someone thought that those discrete numbers are pure science nectar from the gods and thus infallible.

You make it sound like "oh, yes, there was a minor calculation error" when in fact the whole model is rotten to the core.

Again: If you have 5 grades and an error of 1 - do you honestyl think that you can do statistics upon such rotten data? You're assuming here that the basis upon which you base your conclusions is somewhat grounded in reality. When in fact you actually can only roughly guess: Good. Average. Bad.

And then you go and let loos statistics upon this mess and deem it a useful tool, because by damn, those numbers look impressive, and hell yeah, numbers can be used, and godblessus, we don't have to think so much because it's so easy and ohgodjesusyes we can finally pin a goldstar on someone's chest because he's one centrigrade better than his neighbour because the numbers said so!

[Loki]

If you're so obsessed with the ordinal nature of grades, you can just use class rankings. No statistical assumptions are violated (beyond unit homogeneity, which is always violated) if a person's final "grade" is the average of his rank across all of his classes (or across all of his assignments if we're looking for a person's grade in a class).

[Aimless]

Can we think of at least one problem with that approach?