The science
Built on decades of learning science.
Each part of SelfEnabler uses a well-studied way people learn. Here is each one, with the study behind it and a link to the original.
What this research is, and is not
These studies are about the methods: asking before teaching, practising by recalling, correcting a specific mistake, explaining your own reasoning. They are not studies of SelfEnabler. We have no results of our own to report yet.
We built each method into the lessons because the evidence for it is good. Some findings are stronger than others, and further down we say plainly how strong each one is.
Understanding has depth
A single percentage hides how deeply a child understands. So every topic is taught and checked by kind of thinking, up to what the curriculum asks for that class.
The original sorting of the kinds of thinking that school tasks ask for, in six classes.
The revised six steps used today: remember, understand, apply, analyse, evaluate, create.
A short overview of what changed in the revision, by one of its editors.
One lesson, seven steps
Five of the seven steps rest on a named finding. Reading (step 2) and moving on (step 7) follow the order of the ladder itself, so we cite no study for them.
A question first: trying a question before the lesson, even getting it wrong, helps people learn that idea from the lesson. Studied mostly with adults.
How sure are you?: mistakes made with high confidence are more likely to be fixed once the right answer is shown. Studied with adults.
Watch: people learn more from words with pictures than from words alone.
Mayer, R. E. (2001). Multimedia learning. Cambridge University Press.
Watch: ideas held both as words and as images are remembered better. This is a theory, built mainly on memory experiments with adults.
Notes: naming a misconception and explaining why it is wrong changes ideas better than ordinary textbook text.
Check: recalling an idea builds memory better than re-reading it, over days rather than minutes.
Check: children can do more with the right help than alone. Good teaching aims at that gap.
Reflect: students who explained worked examples to themselves understood and solved problems better.
Feedback and help that works
Short, ongoing checks that guide the next step raise learning.
The full review behind it, of about 250 studies.
Feedback works best when it is about the work, not the person.
Bloom argued that almost every learner can master what is taught, given the right help and enough time.
Help for a struggling learner should be different from the first teaching, not a repeat.
Why one-to-one help matters
Bloom's well-known "two sigma" figure came from two small studies. Later, larger reviews found that tutoring helps a great deal, but by much less than two standard deviations. We quote all three, so the figure is never read alone.
In two studies Bloom reported, students tutored one-to-one with mastery methods scored about two standard deviations above students in a normal class.
Across many studies, human tutoring raised results by about 0.79 of a standard deviation: large, but well short of two.
Across tutoring experiments in schools, from pre-school to Class 12, the average gain was 0.37 of a standard deviation.
Taught the way India's curriculum asks
Each stage of school has its own expectations, on a "journey from concrete to abstract".
What each class from 1 to 8 should be able to do, subject by subject.
How strong is this evidence?
Not every finding is equally strong. Here is our plain reading of each, from the kind of studies behind it.
- Recalling instead of re-reading: strong, including with school students.
- Explaining your own reasoning: strong.
- Feedback about the task: strong in direction, though the size of the effect varies a lot.
- Short, ongoing checks: good in direction; how big the effect is, is debated.
- Naming and correcting misconceptions: good, mostly in science topics.
- Words with pictures: good, mostly with university students.
- Mastery learning: good.
- One-to-one tutoring: strong, but much smaller than "two sigma".
- Asking before teaching, and confident mistakes: promising, from lab studies mostly with adults.
- Dual coding, the zone of proximal development and the taxonomies: widely used ideas and frameworks, not tests of a method.
Questions about the research
Has SelfEnabler itself been studied?
No. These studies are about the methods, not about SelfEnabler. We do not have results of our own yet, and we will publish them only when we do.
What is Bloom's 2 sigma problem?
In 1984 Benjamin Bloom reported that students tutored one-to-one did about two standard deviations better than students in a normal class, and asked how group teaching could come close. Later, larger reviews found real but much smaller gains from tutoring.
What is retrieval practice?
Recalling something from memory, for example by answering a question, instead of re-reading it. It builds longer-lasting memory, which is why every lecture has a Dynamic Assessment.
What is the hypercorrection effect?
Mistakes made with high confidence are more likely to be fixed once the right answer is shown. That is why the first question in each lesson asks "How sure are you?"
What is the zone of proximal development?
Vygotsky's name for what a child can do with help but not yet alone. Good teaching aims at that gap.
Does mastery learning mean my child needs 100%?
No. On SelfEnabler, finishing the Dynamic Assessment opens the next lecture. Only "not yet", when a child cannot get started on the basics, keeps it waiting, because that is where a person steps in.
Do you use learning styles?
No. The research does not support teaching to a "visual" or "auditory" learning style, so we do not.
See the science in a real lesson
Book a free demo and watch each step with your child.