A $1 Million XPRIZE is Coming to Education Research

ABOUT THIS EPISODE

Mark Schneider, Director of the Institute of Education Sciences (IES) joined the podcast to discuss the $1 million purse for entrants able to prove the outcomes “promise” of AI and machine learning, and his hopes for what this prize may help spur for educational research culture as a whole.

We're not judging whether or not theintervention that's being tested works. What we're interested in is the abilityof the platform to deliver a rigorous test of an intervention you're listening to enrolment growth,university from helic education, the best professional development podcastfor higher education leaders looking to grow in Roman at their college oruniversity, whether you're looking for fresh and Roman growth techniques andstrategies or tools and resources. You've come to the right place. Let'sget into the show, welcome back to an Roman GrowthUniversity, a proud member of the connect Edu podcast network, I'm ErickOlson with helic education and we're here today with markses director of theInstitute of Education Sciences, Mark Welcome to the show well, thank you forhaving me really excited to have you and talk with you today about the exprize coming to education research. But before we begin, can you get thelisteners a little bit of background on...

...ies and your rule there? Well, I'm thedirector of the Institute of Education Sciences is: Were Science Agency,that's house in the US Department of Education. We are one of the leadinginvestors in education, research in the United States and probably in the worldof it, could be more excited to talk with you about this mark. Maybe to kickUS off today. Many of our listeners are likely familiar with the ex prizedesign to incentive moon, shot innovations literally in some cases,but they may not know that the ex prize is now come to and for education. Canyou tell us about the digital learning challenge expise and how it came to bein me, my pleasure. Yes, everybody knows the ex prize, and actually theremany prize organizations out there. You know that Lily could have gone with,but I think, as you know, everybody knows what the ex prise is encouragesmoon shots literally, and I thought that we could use the x prize in thethe cachet of the expise as a means of...

...spurring on new kinds of innovation ineducation research. So basically, we believe in the strongest evidencepossible and the way in which the gold standard has been set for research israndomized controls, trials, RCTS and I think people probably familiar withRCTS from Spec from medical research, where you have a freaten group in acontrol group in the random patient students, schools are randomly assignedto either those two groups and if the random assignment is correctly is donecorrectly than any observed outcome, differences can be attributed to thetreatment. So this is a very I mean: we've been doing random ize controltrials RCTS in medicine. For you know, for centuries they came to the USeducation system in the strongest way possible in two thousand and two whenis was set up, but the problem with...

RCTS in education is they're hard torecruit students, Thar hard to recruit schools, often leadership changes andthe new leaders say. Why are we doing this? And you know I have otherpriorities, etc and they're long and they're expensive, so they're, usuallyfive years and five million dollars, and I mean even though we're a large inthat during education research. We only do one or two trials on a subject area,and you know at the end of the day, most of the like. Just in dedicanpharmaceuticals, most things don't work, so we've now done one trial on say amasts program doesn't work. We spent five years five. Those hors and thecontrast, of course, is to what's just gone on. With regard to the coidvaccine, I mean that we've done. You know, I mean a lot more money, ofcourse, but we ran multiple trials multiple times and came up with. Youknow the winning hand, but what would have happened if we had done? Oh, let'sjust do fizer for five years. Let's, Oh, that didn't work. Let's do Lodore forfive years: LET'S DO ETCETERA! You...

...can't just do one trial wait for theresults to another Dril, wait for the results, etc. So the digital learningplatform challenge seem to me to be a way to speed up the way in which we doexperiment. So there are many platforms out there. Digital platforms that haveat least a hundred thousand sometimes way more than that the students on them,and so the question was: Can we get the industry that runs these platforms toimagine, doing the equipment of RCTS using the digital platform? So if we dothat, if we, if this works and I'm pretty sure it will by the way what wecould do is we could run multiple experiments across multiple platforms,and what will happen is that some of them will work. Some of them won't workright, but the ones that work. What will then know is it well here. This isa platform that works its delivered, an...

...d experiment. The evidence says thatthis intervention works, but we're going to get that information inrelatively short order and then critically important to the prize isthat you then have to replicate that experiment. So replication is one ofthe most important things that we do in science right, we say: Hey. I work forthis group of people. Does it work for that group of people? I mean we can seethis right now right. We started with above sixty five people. Getting thethe injection and now we're down to well: Does it work for twelve toeighteen year olds, so we're now expanding the populations byreplicating tests across different groups? So this is what so part of theprize is based on the ability of the platform to identify interventions atwork and then replicated at least five and five different populations. So ifit worked in New York, city doesn't work in upstate New York. If it workedin for lowing some black kids in...

Philadelphia, does it work for Hispanickids in Philadelphia, New York? So we need to identify promisinginterventions, then we need to replicate them, because it's only toreplication that we will be able to get the proper fit between what works forwhom, under what conditions? Mark that historical background of the existingconstraints of Education, research that we've been under for so long isincredibly helpful. Higher Education specifically has been talking about theoutcomes promise of technology being able to solve everything. Yea I ammachine. Learning is going to solve everything for years now. I'm curiousfrom your position. Is it the? We have not found that solve? Yet because ofthe constraints you mentioned in terms of the size, the scope, the scale ofthe platforms, we can't learn quickly enough or is it that the technology isnot strong enough yet which do you think it is, and is this x prize goingto be something that helps us to answer that question? Well, actually, what I hope it is. I itis mobilizing industry, so we have...

...close to a hundred companies that havenow put their names into the hopper as being interested in this digitallearning, challenge beautiful, okay, again, almost a hundred companies. Nowmany of them will look at the next step, which is you know now you have to startdoing documentation and you have to start explaining why you should beconsidered for the expise etc and we'll win now it down right. I mean thishappened. Some companies will look at the next wave and just say: Hey Look Iyou know, I don't have the time the money to invest in this gamble. Quitefrankly, so you know some of them will disappear. Maybe we'll get twenty fiveor thirty to get into the next round and then we'll get down to fifteen orso for the final round and then and then we're anticipating about fivefinalists and then the and then the winter so part of what the ex prizedoes is really mobilized in industry right it mobilizes people who are sortof like hey. You know I'm doing a tap form to do business processes acrossschool districts in the United States,...

...but I have a lot of schools, so maybe Icould pivot my platform to run experiments. That's what this is allabout is really trying to mobilize in industry to expand to try the moon shot.So do I think that a I enabled interventions have a promise of course,so we are going to invest in AI, enabled intelligent tutoring, and theseare going to be partnerships. I hope we're still working out the detailswith the National Science Foundation to do to AI institutes, one on intelligent,tutoring and the other one on special education, so not over medicals. Theway we approach Special Ed, but we know that AI has done enormous. has made anormous progress and helping doctors diagnosed problems and identifyingappropriate treatments? And again I have three grandchildren who are invery stages of special education, so I'm very sensitive to the needs and theterms that we use, but the fact the...

...matter is if we could help schools andprofessionals, diagnose, assess special education needs and then come up withappropriate treatments, as a major set forward for about fourteen percent ofAmerican students are have special need. I love those examples. Mark Year seemedexcited about Ai Tutoring, specifically that any other examples of the kinds oftests and experiments that you expect these companies that have put their hatin the ring are going to come with other or that you're hoping to see fromentrance. Well, you know, so we running about three or four differentactivities x prize is the most visible one were science agency, so so we'rejust trying to figure out the most efficient way for science to move intothe modern world, quite frankly for education, science to move into themodern world. So we, and to do that, we obviously need to think about a coupleof things, and that is you know what...

...companies have these the ability tomove into that into the area? How do we support them, but, most importantly,how do we test the Efice of what they, what they're bringing to the table andfor us it's all about F, FC proof of effectiveness, scaling up? Is itcritically as a critical challenge right, so we want these companies notonly to prove that they could work in a in a defined population, but can youscale this up? Can You bring this to scale? Could you affect thousandshundreds of thousands of students and that that is a major challenge andwe're hoping to use our resources to invest more in Scaling Up Yeah? Let'stalk about efficacy, specifically, is a relates to determining a winner here.What does winning look like? What sort of outcomes will it likely take for agroup to win, and how is that winner determined so there's an expert panelthat expise has recruited to actually make the judgment calls I mean I couldtell you what my vision is and they will develop the concrete guidelinesand questions metric to do this. So for...

...me, the winner is one that and againwe're not judging whether or not the intervention that's being tested works.What we're interested in is the ability of the platform to deliver a rigoroustest of an intervention, so this means identifying the treatment and controlgroup or multiple treatment groups right specifically manipulating some.You know, manipulation a strong word and you know like when we don't. Wewant to be careful that we're talking about manipulation of an intervention,not manipulating students right. We so keep that in mind, so we're going todeliver a treatment of math program, for example. We're going to do that.We're going to see how well it works, we're going to what are the metrics bywhich we're measuring how well it works. They have to be standard that metrics.That can't be something that the developer, the program of the thedigital learning platform owner created to have to be a standard metric. Wehave to be. We have to see the capacity...

...of that company to look at the dataanalyze. The data figure out the evidence of effectiveness and then runfive replications. Then, in the ideal world. We don't have enough time to dothis in the ideal world, they would run those replications again right, buteach time you run the replication. You Take the information that you'velearned from the original test and the five replications to identify what thePAFTOR is for further replications. We are not asking them to do the furtherreplications, but they have to show us how they would use the data from thefirst and set around the original test and the five replications to identifythe areas in which are the populations in which they will be doing the nextround of tests. So it's the ability to deliver a test to do the replicationsbased on evidence from the original test plan, but not necessarily exodus.The next round of replications- and all...

...of this is designed to identify to usethe words that we use all the time here and is what works for whom, under whatconditions mark you've spoken, honestly and critically about the traditionallyslow pace of education, educational research, design in this country andsome of the inherent r constraints to make that happen. What are your hopesfor what this prize might do or jump start for educational research, cultureas a whole? Well, we have to be faster, better, cheaper right, you know,there's a joke faster, better to per tell me which to you want. You can't doall three well, I sort of I still hope we coul do all the ice mark. Finally, can you give us a callto arms for those listening excited about this one million dollar purse,perhaps their part of organizations that are interested in participating orthey've a already thrown their hat in the ring? Where should they go next tolearn more? They can contact me...

...directly or they go to the x prize,which has a portal for accepting applications, a call to arms. I thinkwe just did that for twenty minutes. I'm excited I'm ready to bear arms forthis cost great if they're interested in contacting you or your team directlymark, what's the best contact place for them to go. The Mark Dutch neider S, H,N E, I D E R at Ed Dat of awesome, Mark Good luck and thanks so much forjoining us today. It is my pleasure good talking to you by now attractingtoday's new post, traditional learners means adopting new enrollmentstrategies. helic educations data driven enterprise, wide approach toenrollment growth is uniquely helping colleges and universities thrive inthis new education, landscape and Helix has just published the second editionof their enrollment growth play book with fifty percent brand new content onhow institutions can solve today's most pressing enromant growth challengesdownload it today for free at Hillocks...

Education Com play book. You've been listening to enrolmentgrowth university from helic education to ensure that you never miss anepisode subscribe to the show, an I tunes or your favorite podcast player.Thank you so much for listening until next time. I.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (217)