## What We Can Learn About Learning From Khan Academy’s Source Code

December 13th, 2012 by Dan Meyer

It’s great, first of all, that Khan Academy has all their student exercise code on GitHub for everybody to see. I don’t know any other adaptive system that does that. I figured there had to be a better way to reward them for that transparency than the criticism and judgment I’m about to post here, so I made them a badge also.

Their code illustrates the different ways *good math teachers* and *good programmers* try to figure out what students know.

Take proportions, for instance. Here is the code that runs beneath Khan Academy’s proportions assessment.

In each of the dozen files I’ve reviewed, Khan Academy first generates some random numbers that meet certain criteria. In the proportions assessment, they call for three random unique integers between 5 and 12. No decimals. No negatives. No zeroes.

`var numbers = randRangeUnique( 5, 12, 3 );`

Then they use those numbers to generate exercises. With proportions, they insert an “x” randomly into that list of numbers. The final order of that list determines the proportional relationship that students will have to solve.

`numbers.splice( randRange( 0, 3 ), 0, "x" );`

But good teachers are more than random number generators. They create exercise sets that increase in difficulty, that ask students to demonstrate mastery in different contexts, all because proportions are conceptually difficult but procedurally simple. It’s extremely easy for students to get by on an instrumental understanding of proportions alone. (eg. “All you hafta do is multiply the two numbers that are across from each other and divide by the number across from the x.” Boom. It’s badge time.) It’s especially easy when the only thing that changes about the problem is the random numbers.

But forget good teachers for a minute. Let’s look at the bar set by various standard-setting organizations. Here is what you have to do to demonstrate mastery of proportions on a) Khan Academy, b) the California Standards Test, c) the Smarter Balanced Assessment.

**Khan Academy**

You’ll do a handful of problems just like this, with different random numbers in different places.

**California Standards Test**

**Smarter Balanced Assessment**

The difficulty and value of the assessments clearly increases from Khan Academy to the CST and then Smarter Balanced. (I’m hesitant to guess how well a student’s score on the Smarter Balanced Assessment will correlate to all her practice on Khan Academy.)

Here we find a difference between good math teachers and good programmers. The good math teacher can create good math assessments. The good programmer can make things scale globally across the Internet. The two of them seem like a match made in math heaven. Just get them in a room together, right? But the very technology that lets Khan Academy assess hundreds of concepts at global scale — random number generators, string splices, and algorithmically generated hints — has downgraded, perhaps unavoidably, what it means to know math.

**2012 Dec 13**. Peter Collingridge points out in the comments that Khan Academy has a proportions assessment comparable to the California Standards Test. If they have anything similar to the Smarter Balanced Assessment, please let me know.

on 13 Dec 2012 at 8:56 am1 David PattersonKhan is still young and has a lot of kinks to work out. There are big disconnects between the video and assessments. For instance, I looked at Khan’s assessment for graphing linear equations. The very first question I received was to graph -2x=-12. The video Khan suggested to watch if I got stuck ran for 11 minutes and didn’t mention vertical lines at all. Actually, the first sentence in the video was “…any linear equation can be written in the form y is equal to mx plus b”. I sure hope there aren’t students who go to that page as their first experience of graphing linear equations. (My full post about this experience is on my blog.)

on 13 Dec 2012 at 9:09 am2 PeterThere are other proportions exercises on Khan Academy. For example, http://www.khanacademy.org/math/arithmetic/rates-and-ratios/e/writing_proportions looks more similar to the California Standards Test question in your example.

Quite a few of of the Khan Academy exercises do have more wordy problems of the type you mention, and since the exercise module is open, you can always submit some better ones.

on 13 Dec 2012 at 9:11 am3 Kathy Sierra“Boom. It’s badge time.”

Please oh please let this be a meme. *opens Photoshop*

Also, scary post. Necessary, but scary. I say that as a former programmer of old-style CBT. When all you have is code, everything looks like a programming problem. Back then, we believed in the power of computer-based/mediated/assisted learning to deliver customized, personalized “learning experiences”. Which we did. But we forgot the most important thing:

– all that matters is what happens when the clickin’s done

We asked a ton of questions about what *we* (developers, knowledge designers, etc.) should/would do. We asked a few questions about what the “learners” should/would do while “learning”. We asked no questions about what they would do *after* the alleged “learning” took place. And since we had “mastery” neatly packaged in an easy-to-code rule-based system, we had it all covered, right?

Really sad to see the same horribly misguided systems springing up again all these years later, only this time it is different. This time — thanks to the Internet — it’s highly scaled and… global. This time it is unstoppable.

on 13 Dec 2012 at 9:23 am4 Kathy Sierra…yes, this time we have the power of “crowd sourcing” to make things better, but a clear re-read of “Wisdom of Crowds” helps explain how unlikely it is that scaling contributions from the crowd will improve a system that is deeply flawed at the foundation. We can’t A/B test our way out of this.

on 13 Dec 2012 at 9:34 am5 Dan MeyerKathy Sierra:Sounds like ad copy for some kind of terrifying movie. Personally, I’m happy for assessments that ask for better, harder mathematics. I suspect that mastery on KA or any other system that has thought so little about pedagogy will correlate poorly to mastery on the SBAC assessment.

Peter:If you think the defining characteristic of the SBAC task is “wordy,” then we aren’t having the same conversation. More than featuring words, that task asks for application. It asks for the synthesis of several operations. It’s unlikely a student’s teacher has demonstrated a solution to that exact problem type which means it assesses conceptual as well as procedural knowledge.

If you can show me how to add an item like that to KA, I’d be obliged. I’m not sure it’s possible.

on 13 Dec 2012 at 9:36 am6 David PattersonIf you look at the Khan link posted by Peter above, the first two hints for solving the problem look like this:

“There are several equations that could help determine the cost, each with a slightly different approach.”

“Let x represent the unknown cost of 9 erasers. Since 9 erasers cost x, we have the following proportion:

9/x”

What does 9/x signify? The number of erasers per dollar? Is that really the first thing anyone would do when setting this problem up? Doesn’t it make more sense to start with the number of dollars per eraser that we know about? In the example of the California Standards Test above, I’d set up the equation with pounds per acre, not acres per pound. In both the Khan example and the CST example, x ends up being in the denominator in both proportions, which is not the way my brain wants to set them up.

on 13 Dec 2012 at 9:36 am7 B JohnsonGreat post as usual. It drives me nuts how all questions in the assessment piece are identical and at the same level.

A more catchy title for the post might be: Khan Academy refuses to remove the ‘ass’ from ‘assessment’

on 13 Dec 2012 at 10:44 am8 Paul ReimerI think you’re getting at the defining difference mathematical knowledge for teaching makes in designing both learning and assessment items. Take, for example, this item from the University of Michigan’s LMT project: http://bit.ly/UnJBQG.

It’s pretty clear from this item that if you’re not thinking conceptually, any set of fractions will do. But if you’re at all concerned about anything beyond a simple procedure, that set of fractions you throw up on the board for students to compare has just set you back a few minutes of prep time. In their own words, LMT authors created “items we believe represent some of the competencies teachers use in teaching elementary mathematics – representing numbers, interpreting unusual student answers or algorithms, anticipating student difficulties with material.” This list of competencies obviously doesn’t include “generating random exercises for students to solve.”

on 13 Dec 2012 at 11:34 am9 Jason DyerApologies for the tangent, but am I think only one bothered by the Smarter Balanced question tossing “estimated immediate repairs” in the table without any reference in the text? It tests car conceptual knowledge as much as math knowledge.

Also, I would know as a math teacher the problem assumes the price of gas will remain constant over 4 years, but the text just says “the average price in his area” and I could see a student who has too much car conceptual knowledge as having trouble there.

on 13 Dec 2012 at 11:38 am10 Dave MajorWhat I’ve learnt is that I should Github everything I’ve done.

http://github.com/davemajor

I apologise for the most ugly code ever. I hope to release a much cleaner, better, more advanced and more logical boilerplate and tutorial soon.

on 13 Dec 2012 at 11:49 am11 Michael GarciaThe Smarter Balanced assessment is really shaking up the system. Requiring constructed responses from students will fundamentally change how we teach math (which is one of the four types of questions on the new test). Whether through classroom discourse or homework assignments requiring complete sentences, teaching to the SBAC test will not seem so bad.

I would love to see how programming will streamline the free-responses student generate from assessments, but there’s a bigger point here: this new test is promoting a math classroom that requires students to reason and justify their answers via writing. This will be the new normal. Personally, I think this is great and hope will can continue to hold students accountable for their own thinking and writing. I’m not sure how watching videos online is going to help with this.

Our system’s fundamental beliefs about mathematics are implicit in SBAC, and they include that math is about sense-making and sharing our reasoning through communicative means.

on 13 Dec 2012 at 12:51 pm12 James HelakAm I the only one who sees the Smarter Balance Assessment question as having more to do with linear equations than proportions? I believe that you could answer that question without even a basic understanding of proportions. Not only that, but if you set it up as a system, you will see that there is a break even point when Tony drives more than 18,333 miles over the four years. Which car is the better buy is a function of how many miles he drives (all we know is that he dives at least 200 miles per month.) Maybe I’m missing something here, please let me know

on 13 Dec 2012 at 2:26 pm13 Frank NoscheseI blasted through a stack of 8 “wordy” KA proportions exercises in under a minute by following this simple rule:

The answer is the equation with the X opposite the $. If none of the equations have that, chose none of the above.

I even got a “Picking Up Steam” and Lighting Bolts for answering the questions so fast.

You can watch a screencast of it here:

http://www.youtube.com/watch?v=RSCmw8GWdzU

Tell me again how this tests for mastery of proportional reasoning?

(Also, many of KA exercises can be gamed with Wolfram Alpha.)

on 13 Dec 2012 at 2:30 pm14 Gary LogueThe exercises are, according to the video below, on their way to being reworked. Watch from the beginning to 1:24 –

http://www.youtube.com/watch?v=LeeqlXUHNuU

Example Being Referenced

http://sandcastle.khanacademy.org/media/castles/origin:meatglue/exercises/triangle_shadows.html

Patience.

on 13 Dec 2012 at 3:20 pm15 David LippmanAs others have noted, think you set up a bit of a strawman comparing a totally procedural Khan question to application focused questions. But I do acknowledge Frank’s point – that it becomes easy to “game” Khan’s application questions.

So the question becomes: what’s the solution? Random numbers give an endless supply of questions, and work just fine for procedural skills like “factor this trinomial”. But generating an endless supply of unique problems that require conceptual understanding is different. So should Khan and others not even try? Or generate a huge question bank that questions are randomly drawn from?

And a broader question: what is the best way for students to practice conceptual understanding?

on 13 Dec 2012 at 3:54 pm16 DannyOur school started a one-to-one program this year and I was excited to jump into using Khan Academy for my Algebra students. It didn’t last long! I soon realized that many of the practice problems just caused more confusion for my students than they helped. But, I do know that Khan is open for others to add their own exercise sets. I’m excited to see what type of question sets can be created that truly do help with mastery of a concept.

I like the idea of starting with 3 or 4 simple problems as is already in the set and moving up to more advanced problems as it works through. The problem of course is the random generation of questions for the more advanced problem sets. It would require having a set of at least 5 different types of scenarios that random numbers can be generated into. Then the types could be selected in a random order (though they can’t be repeated until all 5 types have shown up).

So a total of maybe 3 stages: KA->CST->SBA with a badge being awarded after a student successfully works through the stages solving 3 in a row at a stage advances to the next stage, missing 2 in a row at a stage drops a stage. Thus a student could gain mastery in 9 questions, but couldn’t just know the basic simple stuff to gain mastery.

Now if only I was a good enough programmer to actually implement that idea! :)

on 13 Dec 2012 at 3:55 pm17 Susan RussoI agree with James Helak – I didn’t use a proportion to solve this problem. Proportional reasoning is an important concept, a useful, real-world concept for even those kids whose mantra is “when will I ever use this in life?” Perhaps a better comparison could be made.

I suppose that’s the problem I have with assessing lots of real-world application problems. I might want students to use a particular technique that we’ve been studying, but there are often much simpler ways to come up with an answer. I debate myself – does the student deserve full marks or not if he hasn’t shown me the math I want but gets the correct answer? I try to tell my students not to reinvent the wheel when working, to use the most efficient method to solve a problem… but then a problem like this comes up.

It’s not so easy, this teaching thing.

on 13 Dec 2012 at 3:55 pm18 Kathy Sierra@David Lippmann

“But generating an endless supply of unique problems that require conceptual understanding is different. So should Khan and others not even try?”

For me personally, it is not what KA (and even more so with some MOOCs) are doing but what they *think* and *claim* they are doing that’s a problem. Use of the word “mastery” is a problem. Reframing the purpose and value of the coded exercises might help.

on 13 Dec 2012 at 4:14 pm19 Dan MeyerDavid Lippman:I amended the post to reflect Khan’s applied proportion problems, which notch it up to the level of the CST. But the SBAC item asks students for more than just an application. As I mentioned earlier:

It has an open text entry field and the prompt to “explain,” fer Pete’s sake.

But don’t let me set up a straw man. Find me a comparable problem on KA.

David Lippman:Best is pretty hard to evaluate, but small class sizes, groups, open tasks, and knowledgable teachers work well. But whenever I blog about KA or MOOCs, I’m not thinking best case. We abandoned best long ago. As I said at the end of this post, I’m okay with less than “best” if we can get “good” to lots and lots of people.

I think

Kathy’scorrect that a lot of online learning’s proponents are under the impression that this is “best,” though.on 13 Dec 2012 at 4:15 pm20 l hodge@Jason & James, you are not the only ones.

@Peter, the better Smarter Balanced questions are actually less wordy, and more mathematical. The example given, is one of the weaker examples I have seen for reasons others have mentioned. Also, doesn’t it seem a little silly to use a proportion for the Khan problem you reference? If 6 rubber stamps cost $7.04 why not just figure out how much one stamp costs & go from there?

All of the eneregy at KA seems to have gone into the data collecting aspect of the platform. If you take that away, it really feels like you are playing with an entry level computer science project – just an excuse to practice some fairly basic programming techniques.

I don’t think you need written answer questions to get useful information, but boy do they have a long way to go in developing useful questions. Really disappointing start.

on 13 Dec 2012 at 4:28 pm21 l hodge@David, yes I would like a large bank of conceptual problems. That would be really helpful.

@Dan, have you looked at the scoring rubric for the SBAC problem? The model answer is just shows the computations. No mention of asumptions about gas prices, miles driven, or anything else. Wouldn’t really call that an explanation.

on 13 Dec 2012 at 4:50 pm22 Kevin H.@Dan, the reason it’s a straw man comparison is that Khan is not just for assessment, but for learning and practice. The challenge is not in creating the questions–I think Khan would have no difficulty writing questions as high-level as SBAC’s. The challenge is that they are committed to having scaffolding that guides students through the solution process and teaches them how to do it if they don’t understand. How do you write scaffolding for the SBAC problem without either giving the whole thing away on one hand, or being so vague as to be useless on the other hand? At this point in conversations about ed tech, you often seem to draw the conclusion that considerations like this are why ed tech is overrated.

That makes me wonder how you’d recommend having students practice questions like the SBAC one. Would you just give it to students on paper and have them work on it in small groups?

on 13 Dec 2012 at 5:07 pm23 Dan MeyerKevin H.Practice is important. Automaticity is important. Conceptual growth is just as important. Khan’s approach to practice and automaticity (randomly generated exercises that are easy to game) undermines our efforts at conceptual understanding. That include Khan’s own videos. Khan could record a lecture that emphasized a conceptual understanding of proportions but if all you need to solve the exercises is for your classmate

Frankto tell you to choose the option that has the x opposite the $ sign, it won’t matter.on 13 Dec 2012 at 6:07 pm24 margretAgreed. Thanks for posting this little revealing fact about Khan Academy. My doc students were spending some time analyzing some student work on four fractions tasks from CMPII that were carefully constructed to elicit different strategies for comparing fractions last week. In short, the numbers you pick matter for which strategy you use. And, teachers probably need to spend more time thinking carefully about what numbers to pick in problems to highlight certain characteristics.

on 13 Dec 2012 at 6:33 pm25 Evan WeinbergAs part of moving to standards based grading this year, I’ve created some problem generators to help me reassess students. I’ve had some success with some Python scripts that add randomized numbers into problems that I’ve written, much along the lines of what KA does with its problems. The biggest issue I’ve found (and that exists with KA) is what to do with the work once it’s done. If the answer is submitted with a click, checked with a click, and badge awarded with little more than several clicks in a row, that isn’t setting the stage for a meaningful assessment of what the student actually knows. In figuring out a student’s skill level, the problem I give them is often just a conversation starter, and the more we talk about their steps and reasoning, the more insight I get into their actual knowledge and understanding.

What I need is a combination of a problem generator/database with a platform for recording and sharing feedback between students and me (or other students). Imagine a BlueHarvest conversation database for each KA problem attempted or submitted as evidence of proficiency in a particular standard. For each problem a student submits an answer to, there would also be a place to submit a screenshot or a webcam photo of the work behind it, and a thread of comments asking for explanation, clarification, or follow up questions.

Any takers?

on 13 Dec 2012 at 7:44 pm26 Leigh NataroIf you see math as a list of rules and algorithms to follow, then maybe KA is for you and your students. And just maybe KA can replace you as a teacher. (Shhhh – don’t let any politicians see this or they may think KA is the way to save $$$$ and save education.)

Creating meaningful lessons and meaningful assessments requires knowing students and responding to how they think on a daily basis. It requires more than a cookbook approach to teaching math. Sure – anyone can follow a recipe, even a computer. It takes a real teacher to be a chef.

on 14 Dec 2012 at 6:20 am27 Bowen KerinsThere’s something else here, too, which is that KA’s approach to mastery and knowledge means you never need to use or apply what you did in Problem #3 to help you solve Problem #4 — and as long as you’ve figured out “the way” to solve them, you’ll just turn the crank.

A good teacher is likely to follow 6/9 = 10/x with 9/6 = x/10 or x/9 = 10/6 or even 9/6 = 10/x, but KA’s algorithms won’t do that.

I’d love to see KA, or anyone else, develop some adaptive problem sets that emphasize connections between topics or sequences of shortcuts, like these:

3x + 5 = 26

3(x-7) + 5 = 26

3(x-100) + 5 = 26

3(2x+1) + 5 = 26

etc.

KA could make sets of fives with these, serving up new numbers each time but the same “point” that whatever’s in the parentheses has to be 7 (in this case), helping kids learn to use variables in better ways.

Come on KA, if Rondo can do it so can you.

http://sports.yahoo.com/blogs/nba-ball-dont-lie/rajon-rondo-taught-algebra-boston-high-school-170112233–nba.html

on 14 Dec 2012 at 8:11 am28 MaryWhat if you’re in elementary school in a class of 25 students, many of which cannot do work a grade level. You have a teacher that’s not a great mathematician (maybe never very good at teaching math) who is trying to get 20 students to understand addition and subtraction. Sure the school says she must differentiate her lessons for ALL students, but she ALL she can do for the advanced students is give them the challenge worksheets. Khan Academy is a real asset for these students. They get immediate feedback on practice problems, rather than corrected worksheets a month later. They have a resource – videos and hints – they can go to when they don’t understand or want to learn more.

My point is NOT ALL teachers, especially in ELEMENTARY schools are GREAT MATHEMATICIANS. Khan Academy takes comments and feedback seriously and improves when they receive it. What if all the great mathematicians put their effort into feedback to Khan Academy, rather than creating critical videos and posting them on YouTube?

Many people have been able to use Khan Academy to get them through their Community College required math class so they can complete a program which gives them a real career. True, they may never understand proportions the way good math teachers would like them to, but that’s not their goal.

on 14 Dec 2012 at 10:23 am29 Kevin H@Dan I agree that poorly-implemented drill problems can make it harder to teach conceptual understanding, if students are internalizing tricks without knowing what they’re doing. My question is: what do you think teachers should do for practice? I use your 3-acts mostly to introduce new concepts, not to practice them. My students’ practice opportunities mostly consist of small-group work on paper-based assignments. Some are simple drill, some are more complex. Is there vastly better stuff out there, especially for non-routine problems like the SBAC question? To me, that’s the need KA could fill, although they’re not that good yet (and as I’ve said before, I think step-based interfaces like Carnigie Learning’s will ultimately win out over answer-based ones like KA).

on 15 Dec 2012 at 4:15 am30 JonI agree with what you’re proposing, but there’s a fundamental flaw in your entire argument: the cost of scoring an assessment of this sort far outweighs the benefits of producing such an assessment. Until federal and state governments pay more than lip service to the improvement of education, easy to score, basal assessments will reign supreme.

Michigan did away with extended response questions because they were too costly. Students hated them because they wouldn’t receive full scores on them if they didn’t answer more than the question asked and did poorly if they answered too much. I still use them in my classroom though because they give me more insight into my students’ problem solving abilities.

Keep on fighting the good fight.

on 15 Dec 2012 at 3:34 pm31 Dan Meyer@

Kevin H, other online outlets like BuzzMath, Aleks, and ST Math have all taken greater care than KA to diversify their exercises beyond random number generation. The tasks themselves change, which gives me greater confidence in the system’s judgment that my students have “mastered” that concept. They are all for-profit systems, I realize. Costshouldbe a consideration here, of course (seeJon’scomment) but it isn’t the only consideration.on 15 Dec 2012 at 6:19 pm32 Kevin H@Dan Thanks, I had never heard if ST Math or BuzzMath.

on 18 Dec 2012 at 9:50 am33 Richard GDan, are there other digital content sites, in addition to ST Math, Aleks, and BuzzMath, that you think are heading in the right direction beyond random number generation?

on 18 Dec 2012 at 10:26 am34 Carl MalartreCreating great content is costlier than we thought. A lot of investors/stakeholders are seeing content as a commodity these days.

@Richard G I like Mathalicious.com’s approach

@Evan Weinberg You are right, collaboration is the future. An app for taking a picture of the detailed work is an idea I’m playing with. I would love to work more on this stuff soon.

@Kevin H Regarding Carnigie Learning’s approach, I’m not convinced. Check this What Works ClearingHouse report:

http://ies.ed.gov/ncee/wwc/interventionreport.aspx?sid=88

If you get your class on BuzzMath, I would love your feedback.

on 18 Dec 2012 at 2:07 pm35 Rolin@mary, your final line is the problem. It’s not about understanding to the point that would make a good math teacher happy, it’s about understanding. Passing a test/bar/competency and understanding what you are doing are two different things. There is a great book by constructivism pioneer Seymour Papert called The Children’s Machine, and it’s about the use of computers in the classroom…while written in 1991, it’s more important today than it was back then. I would suggest the whole book, but there is a section on what Papert calls Kitchen Math that gets to the heart of what it is to regurgitate versus to understand. The beef most (including myself) have with Khan is that it teaches to regurgitation rather than understanding and reproducing in different environments.

on 18 Dec 2012 at 9:26 pm36 MaryThere’s an interesting quote from Papert in a 1984 article Trying to Predict the Future in Popular Computing:

“There won’t be schools in the future…. I think the computer will blow up the school. That is, the school defined as something where there are classes, teachers running exams, people structured in groups by age, following a curriculum-all of that. The whole system is based on a set of structural concepts that are incompatible with the presence of the computer. …But this will happen only in communities of children who have access to computers on a sufficient scale.”

I think Seymour Papert would have seen the potential in Khan Academy.

Khan Academy is a work in progress – it’s not perfect but it gives students access to knowledge and immediate feedback. If these same students should be lucky enough to encounter a really good math teacher, they will be able to improve on their understanding.

To understand Khan Academy, I would recommend Sal Khan’s book The One World Schoolhouse: Education Reimagined. I also suggest looking at Khan Academy by playing the role of student, watching the videos and working the exercises as a struggling student would, following some of the wrong answer paths, with hints and suggested videos. Then take a look at the coach data trail left by this student available to the teacher as coach and imagine how you might help this student.

on 19 Dec 2012 at 10:44 am37 David Wees@Mary

Seymour Papert is still alive, and although I haven’t found any quotes from him on what he has to say about the Khan Academy I think he would agree with, and disagree with you.

1. He has famously said to be patient with early experiments in online learning as they are just that – experiments. This is not the same as abandoning critique of the experiments and provide feedback, but just to remain optimistic about the promise of online learning.

2. On math “flash card” applications:

“Putting flash cards on the screen is not a new way of learning math. It is a polished-up version of the old ways and promotes to greater heights their worst and most mechanical features. Moreover, it is often done in a spirit that I see as dangerously dishonest: Disguising flash cards as a game introduces an element of deception that undermines two educational principles.” ~ Seymour Papert.

The question would be, is the Khan Academy the equivalent of putting up flash cards in the form of videos? If so, then I think I agree with Seymour on this one.

on 19 Dec 2012 at 4:16 pm38 Rolin Moe@David – Papert was injured in a traffic accident a few years back, and has been rehabilitating since…I know of no correspondence on any edu matters from him.

@Mary, I don’t disagree with Khan Academy as a good resource. I have a number of issues with it as a learning system as well as Khan’s educational philosophy. I’m going to try to pinpoint them here as well as reply to your Papert quotation.

1) I have not read all of OWS, but I have viewed Khan’s TED talk, a Forbes interview, and a recent Gates Foundation keynote where he discusses his views on the education system, so I feel like I understand his perspective, I just disagree. He looks at edu as a content delivery system…kids come in, get content, show mastery, matriculate. He doesn’t see why that has to follow the Prussian model of formal ed where students are placed in age-based cohorts. For him, the ease of content delivery in numerous ways (he would call it Personalized Learning) makes our current model ineffective at best.

2) Papert believed in constructivism, a learning theory focused on the student as a maker of knowledge (constructivists often talk of knowledge as something everyone creates) through hands-on practical exercise and authentic experimentation that affect the student and his/her environment (crude summary). For Papert, computers were transcendent machines that allowed numerous opportunities for hands-on practical application that changes environment through programming, design and application. The classroom becomes ineffective here because students can create and program with ease, not be lectured to and run through meaningless exercises of drill and kill.

So while Khan and Papert see the computer as transformational, how they get there could not be more different. Khan Academy replaces the teacher for a video, but it’s still lecture. The flipped classroom says you do your homework at school, but if the homework looks like the quiz problems KA generates in their tutorials, it’s just behaviorist assessment of an inauthentic situation. Papert wanted the computer as a machine of creation, not as a replacement of an outdated learning model.

Khan Academy is not new…it’s technology used to do something we’ve been doing for a long time…lecturing at kids and having them do drills to show a competency. Papert wanted knowledge to transform students…not just competent enough to pass an arbitrary exam, but the ability to use the knowledge in a life context as needed.

I watch how-to videos a lot…how to fold a fitted sheet, how to tie a bow tie, how to change my car oil. It provides me a startin point for a skill I need in that moment (but still need practice for). That being said, tying a bow tie won’t help me launch rockets with my son, and folding a fitted sheet won’t help me communicate a thought effectively. The skills we teach youngsters and want our students to learn are those applicable, transferrable cognitive and creative skills. In my life, I never solve for x on paper…I am presented with a problem and have to utilize algebra in my environment to solve. For some kids, switching from the lecture and worksheet to real life is easy peasy, but most struggle, and in effect don’t do well with math. Khan Academy doesn’t change that paradigm, it just cuts out a weaker lecturer for a stronger one, never questioning that this didactic method might not be the best one.

on 22 Dec 2012 at 5:04 am39 Kevin HKA currently has a 3-6 month contract for a job opening for someone to help them come up with better questions. Looks like it doesn’t involve any programming–just writing questions.

Just thought someone here might be a really good candidate to help them get better.

on 30 Dec 2012 at 2:35 pm40 Mary@Rolin, @David

Sal Khan doesn’t look at education as a “content delivery system”. I did not understand why students must follow an age based cohort system before Sal Khan was born. It goes back to the years I spent sitting in classrooms learning no math because I already knew it, understood it, and had no idea why it was important to learn. Read the chapter in OWS on Mastery Learning, then the chapter on The Economics of Schooling – “The idea is to integrate the technology into how we teach and learn; without meaningful and imaginative integration, technology in the classroom could turn out to be just one more very expensive gimmick.” These expensive gimmick efforts are exactly what Dr. Cuban observes and writes books about. Just as a computer can become an expensive letter opener for a busy executive; it can also become an expensive drill instructor for a struggling student. The goal of integrating Khan Academy into the classroom is not to replace the math teacher but to free up both teacher AND student time to do the real activities that students need in the classroom. These are hands on projects where students can actually use and understand the math skills they have acquired. (See the OWS chapter “Fun and Games”).

It’s interesting that both Seymour Papert and Sal Khan share experience at MIT in common (See the OWS chapter “My Background as a Student”). A discussion between them would prove very interesting. Hopefully Papert will recover enough from his accident at some point in the future to have this discussion. I believe they share a vision of what education ought to be. There is an opportunity to investigate this by comparing Papert’s Logo programming language and the new Khan Academy Computer Science tutorials.

I just finished a month of teaching 5th graders to program using the new Khan Academy Computer Science tutorials, function library, and spin off programs. The classroom teacher and I did this to supplement and reinforce the geometry unit the students were learning. In the midst of this activity we realized that many students never understood ordered pairs until they had to create a program to draw triangles. Computer programming is not part of the elementary school curriculum so we were free to try “flipping the classroom”. That is, let the students go at their own pace, mastering as much programming skill as they were able. What we discovered is the students don’t want to stop learning to program. We have given them access to a tool and they have taken responsibility for their own learning. (See the OWS chapter “How Education Happens”).

Remember Khan Academy is not about watching how-to videos. It’s about wanting to know something and finding a tool that helps you learn.

on 30 Dec 2012 at 6:42 pm41 Technology Gives and Takes Away « Teaching as a dynamic activity[…] From Dan Meyer: […]

on 31 Dec 2012 at 7:11 am42 Leigh Nataro@Mary

I agree that Khan Academy has its uses. If I was a homeschooling parent and wasn’t so great at math, I would probably be using Khan Academy as a resource. Either to re-view the material myself before teaching my child or to have my child watch it directly.

The main criticism I have is that there are people out there (Bill Gates, Dept. of Education gurus) that think Khan Academy can be used as the only resource for learning. They don’t see the benefits of having students work together. They don’t understand that math is more than an algorithm or set of rules. And mainly, they don’t understand the value of a live teacher interacting with a students.

on 31 Dec 2012 at 9:12 am43 Dan MeyerThis comment thread has run its course IMO.

on 24 Mar 2013 at 1:20 pm44 dy/dan » Blog Archive » Why Learning Analytics Aren’t Like Netflix Recommendations[…] BTW. Probably related: What We Can Learn About Learning From Khan Academy’s Source Code. […]

on 01 Jun 2013 at 3:36 pm45 NLB –Post2 When To Practice? | EDCloud[…] take on Khan Academy. But then the mention of Khan Academy and video instruction opens up a whole new debate. Nope, no time for that, need to […]

on 16 Jul 2013 at 4:00 pm46 dy/dan » Blog Archive » What We Can Learn About Learning From Khan Academy’s Source Code, Ctd.[…] Previously: What We Can Learn About Learning From Khan Academy’s Source Code. […]

on 30 Jul 2013 at 12:44 pm47 Lunch Dates: Dan Meyer | wwndtd[…] seem to have a beef with Khan Academy-style stuff (and I happen to agree). But what in particular bugs you about it more than textbooks […]