A thread on Bloom's two sigma problem in education and how advances in technology can help scale effective learning:
1/ Benjamin Bloom in his research found that one on one tutoring can enable students to achieve 2 sigma improvement over conventional teaching.
His objective was to find practical methods that scale to achieve the same two sigma improvement (tutoring isn't scalable).
His objective was to find practical methods that scale to achieve the same two sigma improvement (tutoring isn't scalable).
2/ A brief summary of Bloom's problem:
3/ Why is the two sigma problem important?
Effectiveness of one on one tutoring indicates that aptitude achievement correlation can be drastically reduced. (From 0.6 to 0.25)
It would represent an order of magnitude impact as compared to traditional methods.
Effectiveness of one on one tutoring indicates that aptitude achievement correlation can be drastically reduced. (From 0.6 to 0.25)
It would represent an order of magnitude impact as compared to traditional methods.
4/ How to search for methods to solve the two sigma problem?
Look at the entire teaching/learning process
Break down into mutually independent pieces
Find intervention variables and their effect sizes
Combine variables to achieve intended two sigma improvement
Look at the entire teaching/learning process
Break down into mutually independent pieces
Find intervention variables and their effect sizes
Combine variables to achieve intended two sigma improvement
5/ A brief overview of how Bloom went about searching for these methods
6/ Four categories of sub processes
A - Learner
B - Instructional material
C - Home environment/peer group
D - Teachers and teaching process
A - Learner
B - Instructional material
C - Home environment/peer group
D - Teachers and teaching process
7/ These were the effect sizes associated with each of the variables studied:
8/ Effect sizes by categories:
Most important variables include
Feedback and corrective measures
Cooperative learning
Classroom participation
Tutorial instruction
Reinforcement
Graded homework
Most important variables include
Feedback and corrective measures
Cooperative learning
Classroom participation
Tutorial instruction
Reinforcement
Graded homework
9/ How does technology come into picture here?
Lots of people thought that with widescale access to internet, just making content would be enough.
This led to the MOOC movement which wasn't as impactful as the promise it showed.
google.co.in/amp/s/www.newy…
Lots of people thought that with widescale access to internet, just making content would be enough.
This led to the MOOC movement which wasn't as impactful as the promise it showed.
google.co.in/amp/s/www.newy…
10/ Some reading material on why MOOCs didn't achieve intended results.
spectrum.ieee.org/tech-talk/at-w…
news.ycombinator.com/item?id=165688…
spectrum.ieee.org/tech-talk/at-w…
news.ycombinator.com/item?id=165688…
11/ Basically, MOOCs removed the human element from the teaching process which meant really low completion rstes.
qz.com/65408/the-dirt…
qz.com/65408/the-dirt…
12/ Lots of people who predicted the end of the search for a solution to the two sigma problem were proven incorrect.
This article is a good read on what was expected vs what happened.
Instead of less tutors and less teaching jobs we have more now.
medium.com/@PowerMyLearni…
This article is a good read on what was expected vs what happened.
Instead of less tutors and less teaching jobs we have more now.
medium.com/@PowerMyLearni…
13/ Ed tech today is focused on getting distribution and scale.
Bloom's problem is focused more on significant improvement in summative achievement.
This article explains why ed tech might be heading in the wrong direction.
dataandgraphs.com/analytics/the-…
Bloom's problem is focused more on significant improvement in summative achievement.
This article explains why ed tech might be heading in the wrong direction.
dataandgraphs.com/analytics/the-…
14/ What's next?
Apprenticeship based learning enabled by tech
Access to more people and best teachers, overcome remote work issues, faster feedback and corrective actions, personalized prerequisites work for students, easier self selection of peers.
Apprenticeship based learning enabled by tech
Access to more people and best teachers, overcome remote work issues, faster feedback and corrective actions, personalized prerequisites work for students, easier self selection of peers.
15/ @LambdaSchool has managed to crack this problem by innovating on pedagogy identifying an arbitrage opportunity in the labor market.
The latter is important because it allows lambda to do what they do.
Traditional education has no skin in the game and perverse incentives.
The latter is important because it allows lambda to do what they do.
Traditional education has no skin in the game and perverse incentives.
16/ From my limited reading, a few reasons why lambda school works so well:
- best teachers
- cooperative learning
- hungry students
- mastery learning
- extensive TA support
These variables have high effect sizes.
Sure there's more under the hood that @AustenAllred knows
- best teachers
- cooperative learning
- hungry students
- mastery learning
- extensive TA support
These variables have high effect sizes.
Sure there's more under the hood that @AustenAllred knows
17/ Innovation on the pedagogy front is going to be the moat - it is proprietary and scalable across domains.
@AustenAllred has told us that there's more to come from lambda than just cs grads - datasci, even sales.
Excited to see what's in store.
@AustenAllred has told us that there's more to come from lambda than just cs grads - datasci, even sales.
Excited to see what's in store.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
