Budget crisis be damned, this just seems like a horrendous idea.
I don’t care that an entire class can be “graded” by computer in about a minute while a teacher would take a few hours to do the same work.
I don’t care that it takes $1 to “grade” the paper by computer for every $5 worth of time it takes to grade by a teacher.
I don’t care that newer versions of this program are getting better at “grading” than earlier test runs.
The point is that the computer really isn’t grading. It’s not even coming close.
My Microsoft Word grammar checker puts a green squiggly line under “which” and “that” every single time even if I use them correctly. Online dictionaries often are missing the most basic of words.
Top of the line A.I. programs still can’t hold a conversation that even resembles real life. I mean discussions generally go like this:
“Hi my name is Ender.”
“Hi Ender.”
“What is my name?”
“What is your name.”
I have yet to see a single computer program that can understand logical thought or has the ability to dissect an argument one would use in any well thought out essay.
Every part of this computer grading idea just seems wrong to me. In the long run, the money we save through teachers will be lost because the students weren’t actually being taught anything.
Would anyone care to present an essay detailing the logical counter argument to this? We can even run it by the computer program and see what it thinks.
"HAL: Let me put it this way, Mr. Amer. The 9000 series is the most reliable computer ever made. No 9000 computer has ever made a mistake or distorted information. We are all, by any practical definition of the words, foolproof and incapable of error. "
If they so much as mention using this in Virginia Beach high schools, they will have to deal with a very angry teenager (me, in case anyone didn’t get that :)). I’d send letters to everyone I could think of, talk to my prinicipal, talk to my Gifted Resources teacher, talk to the school board, find out who else could be influential in this matter, send that person a letter, and try to schedule an appointment to talk to that person.
In other words, I do not want my papers graded by a machine. I work long and hard on my papers, perfecting them, honing my writing skills. I make points people would never think of making, take up the side no one else would and prove things from that perspective. I try to look at things from a whole new viewpoint in my papers, to make things informative yet interesting. I like to show things in a new light. And I do not want my efforts to go unappreciated, read only by a machine.
Sorry monica, the Essay Grader 2003[symbol]ä[/symbol] says you must conform and keeps mumbling, “resistance is futile.”
I’d like to see how one of these electronic graders would handle grading the work of a well-regarded but challenging author. I wonder how it would grade the work of Kafka.
Since the want to use this for standardized test grading, I would think a significant side effect would be that teachers would fell pressure to teach essay writing that complies with the grader’s algorithm.
Nah the redwall series by Paul Jaques, all of the animals speak with different accents and it is written with the accents in place as spoken. Oh the horror that a spellchecker would have.
Now for basic spelling and punctuation, let the program have a pass to save time. Content would still have to be evaluated by a teacher for fear of students passing with.
Hello, my name is Fred.
This is my paper.
Do you like my paper?
It is 500 words long.
It is written on white paper.
I spelled everything correctly.
I love to write.
My dog has fleas.
My dog is named Phideaux.
Can you say Phideaux?
I parked my elephant in the red zone.
My lime jello tastes like furniture polish.
In some cases I could see where a hefty percentage of an essay answer could be evaluated via a program.
If a complete essay answer would address several key concepts or people that have specific names it is likely that the answer is basicly correct if those 5-10 key words exist within the answer.
Please note I am referring to tests involving short essay answers. Evaluating a 10 page paper for content would be ludicrous.
Although…with a large enough sample size and some work…hmmm. Take an english teacher and ask 30 students to write a paper on Chaucer. Grade them manually. Give the system samples of “A” papers, “B” papers, “C” papers. after couple hundred papers have chewed though you will have a wide variety of quality papers for the program to compare to. Hundreds of keyword references matched, no spelling, punctuation, or grammar errors caught. Similar balances of prepositions and conjunctions to other “A” papers.
With some work I could see it not being too hard for a fixed subject paper to be graded by machine with a couple years perfecting the program. If a student had a problem with the machine grade and wished to appeal it, the teacher could evaluate it manually.
As a bonus, references could be stored and scanned for outright plagarism in papers without providing appropriate footnotes/cites.
It’s a computer program that will play 20 questions with you.
It features an A.I. that’s been running for 12 years, has played 1,216,304 games and has billions of individual pieces of data from which to draw.
Its success rate is 34.33%. Its success rate with me is 17.74% after 62 games.
This is a fairly sophisticated computer program and I wouldn’t lay even money on it to guess what’s in my pocket. There’s no chance in hell I’d trust it to determine my future no matter how many essays it has as its base of knowledge.
Oh, ick. What an utterly dreadful idea. Students already have to spend the first semester of college unlearning all the horrible writing habits they learn from standardized testing. If their high school training is geared toward learning how to write for computers, rather than people, they’ll be even less prepared.
On the other hand, I like drachillix’ idea of a computerized plagiarism-checker. I’d buy one.
Thats Brian Jaques, unless Brian is an alias.
Electronic essay grading is a bad idea. Computer just aren’t smart enough to recognize good writing. It would be better to just hire a bunch of students to grade papers part time. More expensive than computers yes, but (hopefully) more accurate in grading.
Interesting program. They don’t seem to have acheived linear thinking ability. I chose “a bottle” and it successfully guessed it in 20 questions, but right in the middle, it came up with this complete non-sequitur:
“Does it carry grain in cheek pouches?”
Of course, I established that it was mineral, and hence not an animal in the very first question, so there was no reason to ask that question.
It has a harder time with abstract things. I tried again with “a headache”, and it kept asking things like “Could you use it at school?”. This was after establishing that it was an “abstract” concept. Well, you don’t generally tend to “use” abstract concepts. It never guessed it.
