MI­H­NEA MOLDOVEANU on creative de­struc­tion in higher ed­u­ca­tion

Rotman Management Magazine - - FROM THE EDITOR -

AN UN­PRECE­DENTED — and mas­sively over­due — wave of in­no­va­tion in the higher ed­u­ca­tion in­dus­try is about to be un­leashed, and it will bring un­prece­dented dis­rup­tion to the field.

The waves of dig­i­tal­iza­tion of con­tent, con­nec­tiv­ity and in­ter­ac­tions that have dis­rupted the me­dia, re­tail, travel, en­ter­tain­ment, pub­lish­ing, man­u­fac­tur­ing and fi­nan­cial in­dus­tries are about to strike the higher ed­u­ca­tion in­dus­try, pre­sent­ing a mas­sive op­por­tu­nity for the re­design of a field whose prac­tices have re­mained un­changed since the early 1000s.

This is not a typo: Early Re­nais­sance paint­ings de­pict­ing class­rooms and his­tor­i­cal ac­counts of learn­ing prac­tices both in­di­cate that the ba­sic chore­og­ra­phy of con­tent, con­text, learner-teacher in­ter­ac­tions, and struc­tured drilling and quizzing as a pre­req­ui­site to cer­ti­fi­ca­tion have not changed for more than 1,000 years. The lec­ture-prob­lems-recita­tion-exam for­mat — can­on­ized by re­peated and un­ques­tioned prac­tice in early mod­ern Europe and North Amer­ica — has formed the ba­sis on which learn­ers are sorted, mea­sured, in­cen­tivized, eval­u­ated and ‘taught’.

Re­mark­ably, these prac­tices have per­sisted in spite of a cen­tury’s worth of em­pir­i­cal ev­i­dence — in cog­ni­tive and ap­plied psy­chol­ogy, in ed­u­ca­tional prac­tice, and more re­cently in ar­ti­fi­cial in­tel­li­gence — that there are faster, bet­ter, cheaper ways of help­ing learn­ers ac­quire new skills than those that pop­u­late cur­rent col­lege and univer­sity class­rooms and labs. Spaced learn­ing, vari­able-de­lay re­in­force­ment-based learn­ing, so­cial­ized learn­ing, hy­per-res­o­lu­tion feed­back, prob­lem-based learn­ing — amongst oth­ers — present mod­ern-day ed­u­ca­tors with build­ing blocks for the re­design of the learn­ing ex­pe­ri­ences of learn­ers in ways that in­crease the ef­fi­cacy and ef­fi­ciency of both skill ac­qui­si­tion and skill trans­fer — i.e. the ap­pli­ca­tion of a skill out­side of the con­text in which it is ac­quired.

In spite of the ac­cu­mu­lat­ing and ac­cel­er­at­ing ev­i­dence for the sub-op­ti­mal­ity of cur­rent ped­a­gog­i­cal prac­tices, in­no­va­tion in the ~$3 tril­lion + (2017 USD) higher ed­u­ca­tion field has been slow, spotty and seg­re­gated. The be­havioural blue­prints of learn­ing ex­pe­ri­ences — cour­ses, classes, recita­tions, tu­to­ri­als, quizzes, prob­lem sets, essays, ex­ams — have yet to change in ways that re­sem­ble the re­struc­tura­tion of ev­ery­day ex­pe­ri­ences in the mu­sic, re­tail, pub­lish­ing, travel, com­mu­ni­ca­tions or fi­nan­cial in­dus­tries.

The ex­pla­na­tions usu­ally of­fered for this painful factoid draw on the macro- and mi­cro-in­cen­tives of re­search-ac­tive aca­demics and de­part­ments that use teach­ing--

driven rev­enue to sub­si­dize re­search ac­tiv­i­ties whose out­comes are the ones that ‘count’ and the in­sti­tu­tional forces of re­search-cen­tric uni­ver­si­ties that align in the di­rec­tion of min­i­miz­ing the lo­gis­ti­cal un­pre­dictabil­ity that in­no­va­tion waves trig­ger. They point to the so­ci­ol­ogy and so­cial, cog­ni­tive and de­vel­op­men­tal psy­chol­ogy of homo aca­demi­cus — a crea­ture bet­ter suited and pre­dis­posed to speak­ing about a phe­nom­e­non (say, in­no­va­tion, usu­ally taken to mean in­no­va­tion in a dif­fer­ent field) than to prac­tis­ing it, to anal­y­sis of in­no­va­tive op­tions rather than to the pre­req­ui­site ac­tion and to rep­re­sent­ing rather than to in­ter­ven­ing. Or, they take the ‘tough-mind­edly real­ist’ po­si­tion that higher ed­u­ca­tion is a fil­ter­ing and eval­u­a­tion process of stu­dents for em­ploy­ers, wherein learn­ing and devel­op­ment are de­sir­able but rare and ac­ci­den­tal by-prod­ucts. Of course, this is pre­cisely the sort of (quasi)-causal ex­pla­na­tion whose pro­lif­er­a­tion causally con­trib­utes to the per­pet­u­a­tion of the sta­tus quo.

In the face of such mas­sive syn­er­gis­tic forces, how could it not be that the prac­tice of teach­ing and learn­ing lags behind in­sights and em­pir­i­cal find­ings by a good cen­tury?

But, this ex­pla­na­tion is both in­com­plete in fac­tual base and er­ro­neous in in­fer­ence. The last 10 years have seen mas­sive in­no­va­tion in the field. MIT’S 20-year-old Open Course-Ware ini­tia­tive and Stan­ford’s 30-year-old com­mit­ment to con­tin­u­ous, re­mote learn­ing have mor­phed and pro­lif­er­ated into a mas­sive, open-learn­ing ‘ex­oskele­ton’ which un­der the guises of EDX, Cours­era and Udac­ity bring state-of-the-art con­tent to mil­lions of users while, at the same time, mak­ing it pos­si­ble for ded­i­cated in­struc­tors to learn how to teach from one another.

Cur­ric­u­lar in­no­va­tion in pro­fes­sional pro­grams — notably schools of busi­ness and medicine — has been on the rise since the early 2000s, re­spond­ing to new de­mands for quintessen­tially-hu­man and ex­ec­u­tive skills from ev­er­more-savvy re­cruiters, whose own in-house train­ing pro­grams have also grown in so­phis­ti­ca­tion and size (wit­ness a ten­fold in­crease, from 200 to 3,000, of ‘cor­po­rate uni- ver­si­ties’ be­tween 2004 and 2015).

Re­spond­ing to the need for con­tex­tu­al­ized learn­ing that com­bines con­cep­tu­al­iza­tion and tech­ni­cal skills with the prac­ti­cal know-how pro­vided by con­text, lead­ing-edge En­gi­neer­ing pro­grams — such as the Olin Col­lege of En­gi­neer­ing — have re­designed their learn­ing ve­hi­cles ‘from scratch’, and from first prin­ci­ples, to max­i­mize on the still-elu­sive ob­jec­tive of skill trans­fer from class­room to ‘life’ — and the life-world of or­ga­ni­za­tions, in par­tic­u­lar.

Along­side pos­i­tive ev­i­dence for cur­ric­u­lar and in­sti­tu­tional in­no­va­tion, there is no ev­i­dence that course-level in­no­va­tion hap­pens less fre­quently than does in­no­va­tion in any other field — in­clud­ing those con­sid­ered to be con­sid­er­ably less in­ert than academia. New tech­niques for polling learn­ers, draw­ing them into the so­cial­ized and dis­ci­plined di­a­logue of the class­room, and mak­ing them co-ac­count­able for the ef­fi­ciency of the learn­ing pro­duc­tion func­tion of their pro­gram are find­ing their ways into grad­u­ate and un­der­grad­u­ate cour­ses alike. The cur­rent and bur­geon­ing wave of in­vest­ment in ‘Edtech’ — ed­u­ca­tional tech­nolo­gies meant to in­crease the ef­fec­tive­ness of learn­ing through per­son­al­iza­tion of con­tent and con­text — sug­gests that ped­a­gog­i­cal in­no­va­tion is alive.

Alive, yes; but, why does all of this in­no­va­tion not trans­late into a rad­i­cal trans­for­ma­tion of learn­ing prac­tices across the field? Why does the in­dus­try in­creas­ingly ap­pear to live up to Peter Drucker’s in­dict­ment of it as the ‘largest bur­den on the backs of tax­pay­ers’, stim­u­lat­ing in­creas­ingly-shrill calls for rad­i­cal tech­nol­ogy-based trans­for­ma­tion?

A large part of the an­swer lies in plain sight. Lo­cal, desyn­chro­nized, seg­re­gated in­no­va­tion needs an open, in­te­gra­tive plat­form to gen­er­ate both in­ter­nal mo­men­tum and an in­dus­try-wide trans­for­ma­tion. Ad­vances in telecom­mu­ni­ca­tions — we are now work­ing on 5G sys­tems — pro­vide a telling ex­am­ple. In­no­va­tions in the phys­i­cal and medium ac­cess con­trol lay­ers (lay­ers 1 and 2 of the OSI hi­er­ar­chy) had been fre­quent and sig­nif­i­cant ever since the con­struc­tion of the first dig­i­tal modems in the 1970s. But, un­til the

per­son a liz the the most ef­fec­tive forms of learn­ing are per­son­al­ized to the learner, so­cial­ized to her learn­ing group, and con­tex­tu­al­ized to her work and life.

IEEE stan­dards and the stan­dard set­ting process evolved to the point where in­no­va­tions from com­pa­nies large and small could be syn­chro­nized and har­mo­nized into net­worklevel blue­prints for coding and mod­u­la­tion — i.e. un­til the IEEE cre­ated an in­no­va­tion plat­form — we could not even con­tem­plate send­ing Youtube videos over hand­held de­vices. What IEEE’S plat­forms en­abled for telecom­mu­ni­ca­tions, open-source repos­i­to­ries and plat­forms like Github en­abled for the devel­op­ment of al­go­rith­mic build­ing blocks that took us from Web 1.0 to the cur­rent Web 2.5 — and open AI ini­tia­tives are promis­ing to do for in­no­va­tions in self-re­fin­ing al­go­rithms.

In the ed­u­ca­tional field, the Learn­ing Man­age­ment En­gine (LME) pro­vides the equiv­a­lent in­no­va­tion plat­form that prom­ises to ag­gre­gate and in­te­grate across iso­lated in­no­va­tions in learn­ing and in­struc­tional de­sign. It pro­vides a lo­cus of in­no­va­tion that al­lows both learn­ers and in­struc­tors to learn about the best ways to learn, and to teach while at the same time learn­ing.

A re­cent, large-scale study jointly un­der­taken by the Rot­man School of Man­age­ment and Har­vard Busi­ness School has iden­ti­fied the mas­sive gaps in skills learned and skills trans­ferred that be­sets the higher ed­u­ca­tion field, show­ing that the most ef­fec­tive forms of learn­ing are per­son­al­ized to the learner, so­cial­ized to her learn­ing group, and con­tex­tu­al­ized to her work and life en­vi­ron­ment.

That is pre­cisely what a learner-cen­tric LME will do: It will al­low in­struc­tors to col­lab­o­ra­tively and in­ter­ac­tively de­sign con­tent and learn­ing ex­pe­ri­ences adap­tive to the pref­er­ences, back­grounds, cog­ni­tive and af­fec­tive styles of learn­ers, by in­ter­fac­ing to plat­forms and ap­pli­ca­tions used in re­cruit­ment, ad­mis­sions and alum­nae/i re­la­tions that track learner back­grounds, in­ter­ests and em­ploy­ment pat­terns, while at the same time al­low­ing in­struc­tors to do quick A/B test­ing of con­tent and learn­ing ex­pe­ri­ence de­signs.

Data an­a­lyt­ics — pro­pri­etary and closed in cur­rent sys­tems, but open in the learner-cen­tric LME — will al­low for con­tin­u­ous track­ing of learner pro­files and learn­ing-ori­ented be­hav­iours and learn­ing out­comes, and for in-depth un­der­stand­ing of what-works-for-whom- and-when when it comes to the de­sign of learner-in­struc­tor and learn­er­learner in­ter­ac­tions.

Un­like cur­rent LMES, which do not al­low for instantaneous in-band trans­fers of data be­tween the core en­gine and other learn­ing-en­hanc­ing ap­pli­ca­tions, the learn­er­centric LME will en­able in­struc­tors and learn­ers alike to use and share learn­ing apps in the same learn­ing en­vi­ron­ment, thus deep­en­ing col­lab­o­ra­tion among in­struc­tors and pro­grams and tap­ping into the bur­geon­ing ecosys­tem of ED Tech ap­pli­ca­tions that is cur­rently ‘wait­ing on the side­lines’ and be­ing only spo­rad­i­cally used.

With the flex­i­ble al­lo­ca­tion of de­ci­sion rights to learn­ers and in­struc­tors and the free flow of data and con­tent across pro­grams and schools, a real ‘learn­ing in­no­va­tion ecosys­tem’ will be en­abled. Higher ed­u­ca­tion is a densely and tightly cou­pled net­work of ac­tiv­i­ties and tasks, which in­clude se­lect­ing and mo­ti­vat­ing learn­ers, in­form­ing and test­ing them, con­nect­ing them to in­struc­tors, con­tent and other learn­ers — all while heed­ing the metronome of the aca­demic year and pro­gram guide­lines. If we think of an

LME as a net­work of user be­hav­iours en­joined and en­gen­dered — not just web pages and apps — it be­comes clear that its net­work must have a sim­i­lar level of com­plex­ity to that of the sys­tem it is serv­ing, if it is to func­tion as an in­no­va­tion hub in which in­struc­tors can learn from one another — and from their learn­ers.

Learner-cen­tric LMES will en­able a flex­i­ble al­lo­ca­tion and re-al­lo­ca­tion of de­ci­sion rights over the learn­ing process: Whereas cur­rent LMES give the pre­pon­der­ance of au­thor­ity over class constitution, al­low­able con­tent shar­ing, an­a­lyt­ics, co-hort­ing, apps de­ploy­ment and in­ter­fac­ing to ad­min­is­tra­tors and de­vel­op­ers, a learner-cen­tric LME will al­low in­struc­tors and learn­ers to col­lab­o­rate in the de­sign of the learn­ing ex­pe­ri­ence it­self, by se­lect­ing the ad­di­tional ap­pli­ca­tions, data an­a­lyt­ics, test­ing pro­to­cols and class de­signs that best fit their learn­ing and in­struc­tional ob­jec­tives. Even with­out any of the im­prove­ments in the learn­ing pro­duc­tion func­tion promised — and likely over-promised — by pun­dits and gad­flies in the ‘new AI’ move­ment, proven, re­li­able tech­niques like col­lab­o­ra­tive fil­ter­ing can be de­ployed in the learner-cen­tric LME to pro­duce the so­cial mul­ti­plier of learn­ing ef­fi­cacy that by now has been am­ply doc­u­mented in em­pir­i­cal re­search.

Why Are We Not There Yet?

Once ar­tic­u­lated, a learner-cen­tric LME seems oddly ob­vi­ous as a large piece of the so­lu­tion to fill­ing the ‘in­no­va­tion hole’ of higher ed­u­ca­tion. Why are we not there yet, de­spite hav­ing par-coursed four gen­er­a­tions of learn­ing-man­age­ment sys­tems and en­gines—and of the rat­i­fi­ca­tion of a new stan­dard (Learn­ing Tech­nolo­gies In­ter­op­er­abil­ity) de­signed to as­sure the very kind of open­ness to ap­pli­ca­tions and an­a­lyt­ics we cur­rently lack?

A quick look at the evo­lu­tion of the LMS/LME in­dus­try — cur­rently sit­ting at about US$3.6 bil­lion/year and pro­jected to grow to $7 bil­lion/year by 2020 — gives us the req­ui­site hints. The in­dus­try is heav­ily con­cen­trated around en­trenched providers of ad­min-cen­tric LMS plat­forms that are not in­ter­op­er­a­ble, closed to state-of-the-art an­a­lyt­ics en­gines, and closed to learn­ing ap­pli­ca­tions that are orig­i­nat­ing in the Web 2.0+ en­vi­ron­ment of so­cial­ized, net­work­based learn­ing.

Although some of them started from open source plat­forms ( Can­vas, Moo­dle), they de­vel­oped in­ter­sti­tial mod­ules for data trans­fer and in­ter­op­er­abil­ity that make their cur­rent in­stan­ti­a­tions de facto closed, which al­lows them to charge uni­ver­si­ties richly for an­a­lyt­ics on their own

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.