No US child ever scores less than 40% on a test!

Discussion in 'Science & Society' started by Randwolf, Jun 29, 2009.

  1. Mickmeister Registered Senior Member

    Messages:
    812
    I actually found the article here. I busted out laughing when I read this. LOL...the way we are going, we are going to end up like the ones were in the movie Demolition Man.
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Nasor Valued Senior Member

    Messages:
    6,231
    It makes a certain amount of sense with respect to eliminating statistical outliers. If you have two students who both take 10 tests and one student gets a score of 95% on all 10 while the other gets 95% on 9 of them and a 30% on 1 of them, then statistically there is a high probability that they are in fact of equal ability, even though looking at a simple average of their 10 scores won't reflect that.

    Of course it would be better to actually do a real outlier analysis on a student's set of scores, but we can't expect teachers to actually understand statistics well enough to do that, so simply dropping the lowest score is an approximation. Of course, it would be an even better approximation if you dropped both the lowest score and the highest score...
     
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. iceaura Valued Senior Member

    Messages:
    30,994
    Any statistical method of arriving at a ranked evaluation of the "learning" of a group of students, based on their numerical scores on a few paper and pencil exams, is going to have problems.

    A lot of statistically naive people seem to assume a straight arithmetical mean of the numerical test scores with no adjustments is some kind of more rigorous approach. It isn't - it's just the simplest and bureaucratically easiest technique, the one involving the least work for the evaluators and allowing the least employment of judgment by those on the scene.

    If the exam scores are bounded above and below (0 to 100, say) and the mean scores on the exams overall are well above the median possible (most students score in the 70s on a hundred point exam, say) dropping the low score of a set of exams likely improves the accuracy of a ranking by mean score. Dropping the high score would be less likely to have that effect. That is because an outlying low score has more room to lie out in, and taking the arithmetic mean exaggerates that difference.

    In a five test set with an overall class means of 75 and standard deviations of 10, for example, four scores of 73 and one score of 100 produces a higher rank than four scores of 95 and one of 0. I think most of us would guess that consistently scoring two standard deviations above the mean shows a better relative or comparative grasp of the material than consistently scoring below the mean, eh?
     
    Last edited: Jul 10, 2009
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Nasor Valued Senior Member

    Messages:
    6,231
    Ah, good point, I wasn't considering the fact that these things usually aren't targeted at mean scores of 50%. Given that, it does indeed make more sense to drop only the lowest score.

    Of course, the problem with dropping the lowest score etc. is that it can provide many opportunities for students to "game the system." If there are only a few tests and the lowest one is dropped, then a student who has scored well on the first three tests might reasonably conclude that he has no incentive to learn anything related to the last test. If you are actually doing a rigorous analysis to eliminate outliers then it can get even weirder, since you might have situations where it's better to get an extremely low score than a moderately low score.
     
  8. Randwolf Ignorance killed the cat Valued Senior Member

    Messages:
    4,201
    Well, of course, statistical outliers must be brought into line with statistical norms. What else would we do with the exceptionally bright? (or the exceptionally stupid, but that's a much more unstable political powder keg).

    Statistical outliers point to the extraordinary. Should these people be "normalized"? Or do you have the more benign intention of eliminating "flukes", thereby protecting the "normal" populace? If you would have been in charge of my curriculum, all of my scores would have been classified as statistical anomalies. What then? Eliminate my entire educational record as an "outlier"?

    The truth is, test scores are what they are. Curve them, drop one, divide by the square root of negative one, whatever, the reality is that students were asked x questions (e.g. 3,000) and they answered y right (e.g. 2,500) - therefore they have an 83.3% average.

    Why is this method wrong? How is it skewed? Where is it unfair?
     
  9. Oli Heute der Enteteich... Registered Senior Member

    Messages:
    11,888
    I haven't read most of this thread (so you're allowed to shoot me if my point has already been made) but:
    dropping the lowest scoring mark ensures an "off day" does not affect a student's overall rating.
    That's the reason given when I was on certain courses.
     
  10. DRZion Theoretical Experimentalist Valued Senior Member

    Messages:
    1,046
    There is some merit in this, I suppose.. if you go to school and sit there, bored out of your mind all day, you are still better prepared for getting a real job than if you sit at home and stare at the television. So, getting 40% for just going to school isn't horrendous.

    You have to factor in the fact that simply showing up to school shapes important life habits such as
    -getting up in the morning 5 days a week
    -being more sociable
    -being a tool to the government
    -learning about life, and your peers
    -learning your place in society
    -etc. etc. etc.

    And I think a person who goes through the hell that school can be deserves some kind of 'reward' because shit, they ain't there for free.
     
  11. Baron Max Registered Senior Member

    Messages:
    23,053
    Why don't we just give the little bastards whatever grade they want and be done with it? Why bother grading if the grades don't mean jack-shit to anyone? So, ...at the end of the school year, just let the little fuckers pick whatever grade they want for each subject. Fuck it, who cares? If we don't care and the little bastards don't care, why should anyone else care?

    Besides, the nation is becoming socialistic anyway, so the little bastards don't even need to work ....the gov will hand out money, gvie them housing and food. Fuck, why do anything?

    Baron Max
     
  12. pjdude1219 The biscuit has risen Valued Senior Member

    Messages:
    16,479
    So how would you explain those socialist states doing better than the US academically?
     
  13. ripleofdeath Registered Senior Member

    Messages:
    2,762

    as long as you dont shit yourself then roll around on the floor in it and call that school then it is easy to do better than any american school.
    check out their adult illiteracy rates.
    it should be a crime for the state to fail soo miserably in something they are tasked to do by an act of law and mutual contractual agreement.

    http://www.proliteracy.org/NetCommunity/Page.aspx?pid=335&srcid=393

    Basic Facts about Literacy

    * Literacy is the ability to read, write, compute, and use technology at a level that enables an individual to reach his or her full potential as a parent, employee, and community member.
    * There are 774 million adults around the world who are illiterate in their native languages.
    * Two-thirds of the world’s illiterate adults are women.
    * In the U.S., 30 million people over age 16 — 14 percent of the country’s adult population — don’t read well enough to understand a newspaper story written at the eighth grade level or fill out a job application.
    * The United States ranks fifth on adult literacy skills when compared to other industrialized nations.
    * Adult low literacy can be connected to almost every socio-economic issue in the United States:
    o More than 60 percent of all state and federal corrections inmates can barely read and write.
    o Low health literacy costs between $106 billion and $238 billion each year in the U.S. — 7 to 17 percent of all annual personal health care spending.
    o Low literacy’s effects cost the U.S. $225 billion or more each year in non-productivity in the workforce, crime, and loss of tax revenue due to unemployment.
    * Globally, illiteracy can be linked to:
    o Gender abuse, including female infanticide and female circumcision
    o Extreme poverty (earning less than $1/day)
    o High infant mortality and the spread of HIV/Aids, malaria, and other preventable infectious diseases


    o More than 60 percent of all state and federal corrections inmates can barely read and write.

    the state failed them as a child then fail them again in a second chance in rehabilitation.

    unles your diagnosed as a psychiatric patient you should be required to read and write before you leave prison.
    if you cant then you are sent to a low security school in the middle of no where, where you are schooled in basic life skills reading writing and how to apply for jobs, take care of a house and manage its bills and basic child care if you have children.consumer rights, basic small management practices and basic cooking with government funded options to carry on and do things like apprenticeships.

    i bet the reoffending would be slashed by 50%.

    and not released until you pass.
     
    Last edited: Jul 11, 2009
  14. Meursalt Comatose Registered Senior Member

    Messages:
    395
    Ever occur to you to wonder why we can't?
     
  15. iceaura Valued Senior Member

    Messages:
    30,994
    I'm just as happy giving a high grade to a student capable of gaming such a system, if reasonably set up, as I am giving a high grade to a student who has rote regurgitated in calculated patterns without comprehension.
     
  16. Baron Max Registered Senior Member

    Messages:
    23,053
    Because in those socialist nations, the grading system is kept secret so they can spread the applicable propaganda claim their kids are better than our kids.

    Pick something, most anything, any form of business or inventions, etc that you'd like ...check out how well Americans do in that. Americans lead the pack in almost every field or business there is. Perhaps Americans do better than the grades would suggest, huh?

    Fuck all them thar furriners!

    Baron Max
     
  17. Saven Registered Member

    Messages:
    209
    40% is still a failing grade so what difference does it make, except to make it so missing one assigment doesn't utterly destroy your overall grade forever.
     
  18. Baron Max Registered Senior Member

    Messages:
    23,053
    Don't worry ....that's next!! In a few weeks, the US school systems will enact a new rule that no kid should ever get a failing grade in any subject and shall never be held back to learn anything.

    Baron Max
     
  19. pjdude1219 The biscuit has risen Valued Senior Member

    Messages:
    16,479
    Now your just being paranoid and a conspiracy theorist.
     
  20. ripleofdeath Registered Senior Member

    Messages:
    2,762
    that's why i could never work in the school system.
    i just cant stand the apathy toward learning and people that requires you to just stand by and watch people fall through the cracks.

    all the false statements, lies and crutches that would be needed to try and feel self worth for knowing you were failing children over and over year in year out.

    while i may be somewhat guilty of giving too much value to some human life, i am just as guilty of giving no value to other human life.

    i have rarely(not entirely sure i have ever) ever met a child that was worthy of a life sentence.


    i suspect what i have written above is one of the biggest reasons women like early childhood education soo much, because they cant perceive any real damage that has been done by bad teaching and bad interaction by the school/teacher.

    and women rule the living in denial thing hands down.
    although they are starting to draw near even with men now.

    P.S as women pull even with men in being real and not playing happy families(etc...) then men may well have a surge in domestic violence possibly other violent crime suicide included.
    but it will just be a bubble as long as it is stamped on by police and society.
     
  21. Randwolf Ignorance killed the cat Valued Senior Member

    Messages:
    4,201
    I think maybe you are misinterpreting the OP. The idea being batted about is whether or not a child who does not go to school on exam day deserves a 40% grade. The intent was not explicitly nor implicitly designed to be a method for rewarding children for attending school, rather more nearly the opposite effect.

    I believe someone already pointed out a hypothetical case in which a student is cruising along with, for example, a 90% average and an exam is scheduled for tomorrow. Said student knows or (highly suspects) that they will score less than 40% on this exam, because they haven't studied the material or were unable to grasp the content for whatever reason. They now have an incentive to skip school that day, knowing that they will "earn" a 40% grade on the exam for not showing up. This is good?

    As to the statistical guru's out there, doesn't the appropriate grading / analysis system depend on what we are intending to measure? In other words, if you are trying to plot the distribution of the results of various tests based on equality, then a more complicated system involving standard deviations, outliers, ranges and modes might be appropriate.

    On the other hand, if we just want to know that little Johnny and Mary were both administered 2,000 questions throughout a course's duration and Mary got 1,940 of the answers correct, while Johnny got 1,760 correct, then it would seem that a straight up right/wrong percentage would be the more accurate method, i.e Mary's grade is 97% and Johnny's grade is 88%.

    As to the "bad day" effect, I agree that it was probably the original intent of dropping the lowest score to offset this phenomenon. As I stated earlier, I didn't really intend to focus on the "drop one" practice, it seemed merely to be the beginning of a "slippery slope". I also pointed out that it at least applies to all students, not a certain subset.

    This is my main problem with the proposed "40% system". It is not applied equally, student "A" with perfect attendance that scores 38% on a particular test will end up with a worse overall score than student "B". This is based on student B "playing the game" properly and skipping school on the day that he / she would have scored 10%. If student A and student B had an otherwise identical record, then we have a situation in which we reward student B for not taking a test at all.

    If we are comfortable rewarding kids who excel at working the system, finding innovative ways to cheat (like the cell phone texting scandal a few years back), etc., well alright then. Let's at least just call it what it is - skewed results, no matter how good the intentions or motives are.

    I know the educational system fawns over "new math", but no way in hell does 40 = 0. Go ahead, take a day off from work, you will still get 40% of your normal pay... :shrug:
     
  22. Nasor Valued Senior Member

    Messages:
    6,231
    For god's sake, I was talking about statistical outliers with respect to that individual student's set of scores, not with respect to all student's scores. What you seem to be imagining (considering an entire student to be an outlier) would only be useful if you wanted to do something like rank different teachers against each other by comparing the scores of their students.
    Really? Every single score you got was a statistical outlier with respect to all your other scores? I would be very interested to hear what your scores were, because that would be mathematically impossible.
    Because, as was already explained, it doesn't account for statistical outliers. I don't really feel like writing a long intro to statistics lesson on what an outlier is or how outliers can skew data when you have a limited sampling (like, say, a student who takes a limited number of tests). But if your task is to rank a bunch of students with a limited number of tests, your rankings will generally be more accurate if you eliminate statistical outliers. Always dropping the lowest score isn't the same as actually doing a statistical analysis of the grades to identify outliers, but it has much the same effect.
    Well shit, now you're just making me annoyed. The concepts of standard deviation and outliers don't have anything to do with "equality," they have to do with analyzing statistical data so as to derive the most accurate conclusions. Go take an intro to statistics class at your local community college or something.
    Oh, so we're only concerned with who got what scores? In that case no statistics are needed; it's a matter of simple arithmetic. But if you're interested in trying to rank students according to their ability, then employing statistics to analyze those scores can indeed make your rankings more accurate than if your just take the raw data at face value.
     
    Last edited: Jul 12, 2009
  23. Nasor Valued Senior Member

    Messages:
    6,231
    Perhaps if the class is on game theory, or math, or economics. But it seems stupid to reward someone for gaming the system in, say, history class.
     

Share This Page