This is probably a repeat comment from me. A two-sigma ratio in knowledge is almost 4.5. the standard US grade level to grade level improvement of the median student is about 1.7 standard deviations (SD or "sigma"). a knowledge ratio of about 3.8. This is related to relative knowledge, not absolute knowledge. If you plot two Gaussians side by side, you will see that 1.7 SD steps are quite ideal for mass teaching, because overlaps of higher performers in the lower grade class to higher performers in the upper grade level intuitively are what one would like to see. If every student was motivated, and if every teacher was motivated, and if every classroom was NOT populated with the full spread of +/- 2.0 sigma, then everyone would be learning more absolute knowledge. The average 5th grader might know as much as the 7th or 8th grader of years before. The grade to grade separation would still be 1.7 SD, but the new SD would correspond to more absolute knowledge. ... Bill Lynch ... bandglynch@gmail.com
You are spot on. The early American frontier schoolhouse had students teaching students and progress made at their own pace, and not based on age and simply being pushed forward. Further, Carolyn Dweck (Stanford) helps us see that intelligence is developable and much of this prospect revolves around good curriculum.
Marty: TY. This is a complex matter with passionate people on both sides. It seems to me that if we are to have any hope of fixing our broken K-12 education system, we should be able to have an open-minded, civilized discussion of the pros and cons of our education options.
This may be somewhat of a side note, but I personally observed a "C" to "A" change in the college work of a friend about 50 years ago. She signed up for Dick Pothier's "Journalism 101 - Newswriting" class at Northeastern University in Boston. Dick's method was that each student must write a coherent news story out of a jumble of presented facts. It's the same activity that every good newswriter does every day.
There were two parts to the assignments: one in class and one as homework. The work was timed--ten minutes and you're done, just like at a newspaper with deadlines. "Here's a fact sheet. Write a story." The work included discovering the lede and arranging the facts correctly. There was no excuse for not doing the homework. "If it takes you longer than ten minutes you're not doing it right." Those assignments were 90% of the student's grade.
My friend found that learning the focus needed for that one course improved all of her grades in all subjects.
Dick died young, only 55 years old. Here's his obit.
To this day I enjoy reading the sports news in newspapers. I'm not a sports fan at all, but I love to see if the writer has captured the lede. Some do it well, some not.
John…I’m breaking my promise to you that I would not read your Bloom critique 🤪.
As I’ve told you many times, my opinion is that the Bloom study is intuitively obvious, supported by his experiment…that is…one-on-one tutoring with adjustments based on mastery monitoring is a far superior pedagogy to a teacher in front of a class of wildly varying competency…basically teaching at grade level instead of teaching at knowledge level. Doing this at a mass level (50M students) was not practical or affordable in 1984 (Bloom’s analysis), but is now practical, scalable and will quickly become affordable.
You mistakenly said that current AI platforms, such as Alpha Schools, are discarding conventional “direct instruction” pedagogy. Some education platforms use “direct instruction”, such as Thales. However, the vast majority of public, private, charter education use classic teacher lecturing at grade level with little mastery monitoring. This system is simply inferior, and our deplorable results across America prove it. No state has demonstrated superior K-12 education performance.
It is far past time to have a scaleable pedagogy with correct curricula with demonstrated superior performance applied to our 50 million students. We now have the technology and certainly the financial resources (we spend $1 trillion/year on a failed system) to get this done.
I believe that the use of AI in education is still actually at an infant stage. (I pray) What I have found the most appealing with current learning models is the ability to change the AI mind (sic). To conceptualize a question to make the AI consider an alternative to original answer requires incredible Critical Thinking…
…it would be this type of expercise that may prove extremely valuable in improving student outcomes in that area
Like a debate team. You have to be able to debate EITHER side of an argument through logically deduced questioning
I disagree that this is the case - “Artificial Intelligence (AI) models excel at drills, explanations, and pattern recognition but still struggle with teaching critical thinking, creativity, and deep conceptual understanding.” It depends on how it is used. For sure, if you just ask ChatGPT or Claude for information or even to help you learn, you will likely get this basic approach but it is not difficult now to reframe the AI to output something closer to what you suggest it is not capable of - teaching critical thinking, creativity and conceptual understanding. We are trying to achieve exactly this with Education Hero and I think I made quite a bit of progress with my weekend project too - https://www.socratic.institute/. The pace at which AI is developing and people harnessing its power in ethical and appropriate ways means that it will become optimal in most aspects of pedagogy sooner rather than later, IMO.
This is probably a repeat comment from me. A two-sigma ratio in knowledge is almost 4.5. the standard US grade level to grade level improvement of the median student is about 1.7 standard deviations (SD or "sigma"). a knowledge ratio of about 3.8. This is related to relative knowledge, not absolute knowledge. If you plot two Gaussians side by side, you will see that 1.7 SD steps are quite ideal for mass teaching, because overlaps of higher performers in the lower grade class to higher performers in the upper grade level intuitively are what one would like to see. If every student was motivated, and if every teacher was motivated, and if every classroom was NOT populated with the full spread of +/- 2.0 sigma, then everyone would be learning more absolute knowledge. The average 5th grader might know as much as the 7th or 8th grader of years before. The grade to grade separation would still be 1.7 SD, but the new SD would correspond to more absolute knowledge. ... Bill Lynch ... bandglynch@gmail.com
You are spot on. The early American frontier schoolhouse had students teaching students and progress made at their own pace, and not based on age and simply being pushed forward. Further, Carolyn Dweck (Stanford) helps us see that intelligence is developable and much of this prospect revolves around good curriculum.
Marty: TY. This is a complex matter with passionate people on both sides. It seems to me that if we are to have any hope of fixing our broken K-12 education system, we should be able to have an open-minded, civilized discussion of the pros and cons of our education options.
You are correct that there are many views on "the fix." Keep pushing forward the importance of Critical thinking and much will take care of itself.
This may be somewhat of a side note, but I personally observed a "C" to "A" change in the college work of a friend about 50 years ago. She signed up for Dick Pothier's "Journalism 101 - Newswriting" class at Northeastern University in Boston. Dick's method was that each student must write a coherent news story out of a jumble of presented facts. It's the same activity that every good newswriter does every day.
There were two parts to the assignments: one in class and one as homework. The work was timed--ten minutes and you're done, just like at a newspaper with deadlines. "Here's a fact sheet. Write a story." The work included discovering the lede and arranging the facts correctly. There was no excuse for not doing the homework. "If it takes you longer than ten minutes you're not doing it right." Those assignments were 90% of the student's grade.
My friend found that learning the focus needed for that one course improved all of her grades in all subjects.
Dick died young, only 55 years old. Here's his obit.
https://www.findagrave.com/memorial/90505788/richard-joseph-pothier
To this day I enjoy reading the sports news in newspapers. I'm not a sports fan at all, but I love to see if the writer has captured the lede. Some do it well, some not.
Peter: Thank you for sharing that personal story!
John…I’m breaking my promise to you that I would not read your Bloom critique 🤪.
As I’ve told you many times, my opinion is that the Bloom study is intuitively obvious, supported by his experiment…that is…one-on-one tutoring with adjustments based on mastery monitoring is a far superior pedagogy to a teacher in front of a class of wildly varying competency…basically teaching at grade level instead of teaching at knowledge level. Doing this at a mass level (50M students) was not practical or affordable in 1984 (Bloom’s analysis), but is now practical, scalable and will quickly become affordable.
You mistakenly said that current AI platforms, such as Alpha Schools, are discarding conventional “direct instruction” pedagogy. Some education platforms use “direct instruction”, such as Thales. However, the vast majority of public, private, charter education use classic teacher lecturing at grade level with little mastery monitoring. This system is simply inferior, and our deplorable results across America prove it. No state has demonstrated superior K-12 education performance.
It is far past time to have a scaleable pedagogy with correct curricula with demonstrated superior performance applied to our 50 million students. We now have the technology and certainly the financial resources (we spend $1 trillion/year on a failed system) to get this done.
Peace!
Don: TY for perusing my commentary, but your comments did not address any of the critiques I listed.
You also seem to blame our inferior results on pedagogy, where the evidence is that the blame is primarily on the deplorable curricula.
I believe that the use of AI in education is still actually at an infant stage. (I pray) What I have found the most appealing with current learning models is the ability to change the AI mind (sic). To conceptualize a question to make the AI consider an alternative to original answer requires incredible Critical Thinking…
…it would be this type of expercise that may prove extremely valuable in improving student outcomes in that area
Like a debate team. You have to be able to debate EITHER side of an argument through logically deduced questioning
Peter: Yes, relatively speaking, AI might be considered to be an an infant stage.
Some infants grow up to be competent, productive adults. Others end up being evil.
Amen✝️
Maybe “thoughts and prayers” is a better option or at least an adjunct teaching…
I disagree that this is the case - “Artificial Intelligence (AI) models excel at drills, explanations, and pattern recognition but still struggle with teaching critical thinking, creativity, and deep conceptual understanding.” It depends on how it is used. For sure, if you just ask ChatGPT or Claude for information or even to help you learn, you will likely get this basic approach but it is not difficult now to reframe the AI to output something closer to what you suggest it is not capable of - teaching critical thinking, creativity and conceptual understanding. We are trying to achieve exactly this with Education Hero and I think I made quite a bit of progress with my weekend project too - https://www.socratic.institute/. The pace at which AI is developing and people harnessing its power in ethical and appropriate ways means that it will become optimal in most aspects of pedagogy sooner rather than later, IMO.
Joel: We don't know for sure what the future holds, but that quote was about how an education person saw things now.
As I wrote, there is great promise for AI, but also extraordinary hurdles to overcome.